How would Google farm?

Abstract

If Google were to enter farming, it would likely start by finding and amassing all available data – without biases or preconceived notions – and let the data dictate how to make on-farm management decisions. Google’s self-driving car can see, hear, read, understand, decide, and act – just like human drivers do. We now also have tremendous data collection capabilities for situational awareness of the crops we are growing. As Google revolutionizes how we think about mobility, we have the same opportunity to exploit big data in agriculture. We are learning how to combine the latest advances in precision agriculture and plant genetics with big data and advanced artificial intelligence. As the world’s population continues to grow rapidly, merging these technologies is critical to maintaining an adequate supply of food and fiber.

How can ‘big data’ transform smallholder farmers’ lives and livelihoods?

Abstract

During the life of this generation, the world will have to overcome one of its greatest challenges ever faced by humanity: Within the next 30 years, the earth’s population will increase by 50%, increase that will be most visible in Africa and Asia. In order to have a world fully and well fed before 2050, we will have to increase food availability by more than 60%; this means that this century alone the world will have to produce as much food as it did over the last 8000 years. What is the solution to this problem? Innovation; innovation through data. Data is empowerment; it allows people to make better, wiser, more efficient choices. Whether it is consumers buying more nutritious food, hence increasing the demand for it, producers responding accordingly, and more efficiently; transporters reducing costs because they can better know what will be available where, and when. Governments are also better empowered, more able to predict where climate change will require production changes, increasing trade and services to the population.

However while a relatively high level of productivity has already been achieved in large farms across the developed economies, most of the untapped agricultural potential that remains untapped lies with the small farm holders. We know that while economies of scale did play an important role in improving productivity on the large farms, data, and in particular its precision agriculture and irrigation also played a key role. Yet until recently small farm holders, already at a competitive disadvantage because of the small size of their plots, saw this compounded by the fact that data was the privilege of the few, of the large organisations that were the only ones able to afford high level technicians and expensive equipment required to process and understand big data….This has changed.

New economic models have emerged through which it is cost effective for Governments, as well as for private sector, to open big data to allow it to reach small farmers, traders, and consumers alike. Technology meanwhile made giant leaps allowing a new generation of software and hardware tools allowing non experts to make use of big data in their day to day management of their agriculture/food business, even in the most remote parts of the world. In this presentation we will look at how access to big data initiatives taken by groups of individuals, enterprises (big and small), governments, are taking up the global food security challenge ahead of the world, while changing the lives of small farmers, through efficiency, empowerment and better quality of life for all.

The use of Big data in precision agriculture

Abstract

Precision agriculture (PA) tools, technologies, and information management strategies are becoming mainstream in many regions of the world. Several technologies such as automated guidance, yield monitoring, and georeferenced soil sampling are no longer considered premium services, but are standard offerings from many agricultural providers. Adoption of these types of technologies that don’t depend on site-specific information to extract value has been rapid and steady, while others that require agronomic calibration or some level of data management such as remote sensing, soil management zone delineation, and using yield maps to characterize and understand field variability has been slower and a much greater challenge than many would have predicted two decades ago. Information needed for farm management is also more accessible and delivered faster through mobile device technology than ever before. This increased access to information is driving a digital revolution in production agriculture. The key for the industry is to develop strategies to effectively utilize the abundance of available data to make better decisions on the farm. The term “big data” typically refers to a volume of data too large to be handled by traditional processing platforms. So, what counts as “big data” depends on the context. In precision agriculture, “big data” can include spatially referenced and temporal data. There are varying scales of use being evaluated such as localized on-farm data, pooled regional data, and more traditional “big data” sets that are more comprehensive but can be mined to address specific local conditions. Despite the lower costs and much greater capacity to store data compared to twenty years ago, the information side of PA has struggled somewhat in demonstrating value. The objective of this presentation is to present examples of big data in PA and to discuss current industry and academic efforts to overcome existing barriers to using data to better serve the farmer.

Read More

Opportunities for maximizing crop productivity – insights from agronomic data : Use cases from Kenya and Tanzania

Abstract

Increasingly, data from different sources are becoming available. While these data contain valuable information, conventional methods of data analysis are largely unable to handle such data.

Application of machine learning techniques help to unravel patterns for further attention. Three approaches are presented 1) boundary line analysis applied to derive application rates of P for western region of Kenya; 2) application of regression trees to inform landscape-level management using 2 examples from Tanzania and 3) cluster analysis to understand response patterns and their drivers using multi-locational omission trials data from 5 countries. Machine learning approaches offer huge potential for application in agronomy.

Read More

Precision Farming Services, Drones, Satellites and state of the art technology for agriculture

Abstract

The quick evolution of the technology in the last years have allowed the farmers to have many tools and different techniques that where impossible to obtain just one decade ago.

One of these state of the art technologies, and which is evolving at a high speed, is the use of imagery which is known as DRONES, but that is also now available through satellites and for specific cases through land sensors which are hand manipulated or mounted on vehicles.
Through specific images of DRONES and satellites obtained in different fields in the African Continent we will present the results generated by both, and the way these are applied and interpreted for the use in agriculture.

This analysis will take us from the DROENS and Satellites down to the soil and the properties of this, in addition to the nutrients and plant tissue analysis which by combining all of them we can understand what the images and colors finally tell us about our crops.

These images, which represent maps of the field will be also further discussed including different kinds of maps such as soil fertility, yield and even nematode ones. The use of specific software for this kind of technology and it´s practical use and importance will be part of the discussion.

Read More

Data from airborne imagery. Barriers and advances for an efficient application in agri-environmental engineering

Abstract

Increasingly complex issues in ecological engineering, land-use planning and precision agriculture make it obvious that aerial data acquisition is needed to satisfy large-scale mapping and spatialized monitoring. For decades, airborne and satellite imagery has been used as a means of decision support. Indeed, these means of acquisition allow mapping of large areas and enrichment of this cartography with contents making it possible, among others, to plan interventions, to map risks, even to elaborate scenarios of managements …

This conference aims to present recent advances in the development of innovative approaches in the acquisition of aerial images of vegetation and the treatment of these images based on semi-supervised segmentation that allows to process data acquired on important surfaces. These innovative approaches coupled with machine learning make it possible to develop cartographic data that are increasingly reliable at large scales and with the least recourse to expert intervention. Examples of applications of these technologies in the field of agri-environmental engineering will then be presented.

The final part of the conference will focus on the contribution of the systemic approach to the efficient use of these technologies for the development of functional and dynamic information systems for ecosystem management, with a focus on the place that should be taken by soils in these systems.

Read More

Information systems for decision support in agriculture

Abstract

INRA-Morocco developed inovative decision support tools for agriculture, thanks to key methodological and technological achievements since more than 20 years continuous research. Major tools include the development of the first modelling system called MOSAICC dedicated to the impact of climate change on agriculture, water and forests sectors, developed as part of the EU / FAO program on global governance and hunger reduction of FAO coordinated by INRA. In terms of climate risk management in agriculture, INRA developed a crop monitoring and crop yield forecasting tool called CGMS that allows the monitoring the agricultural season based on meteorological and satelite information. INRA also developed a decision support system for the fertilization of major crops in Morocco, thanks to a research & development project initiatied by OCP-Group and the Ministry of Agriculture.

Read More

Artificial Intelligence and Big data for impacting small holder africain farmer livelihoods

Abstract

Artificial Intelligence and data science offer unique opportunities to transform alternate data sources and obtain interesting solutions particularly to address problems in low-resourced countries like those in Africa. A key problem is the lack of experts across all fields. In this talk I will highlight some of the examples of interesting problems that can be solved by computational techniques, the unique and alternate (Big) data sources that can be leveraged to address these problems and the lessons we have learned in doing this. A common solution to most of these problems is to automate a critical expert task. Some of the topics I will cover will be in Agriculture, e.g. automating the diagnosis task using images and crowdsourcing surveillance data from small-holder farmers. Most of the work we do is targeted towards a particular client so I will also highlight some of the experiences we have had dealing with the consumers of these technologies.

Read More

Data-driven approaches for transforming agriculture in Africa – Agrimetrics case

Abstract

This presentation will first provide an overview of the work of Agrimetrics, the world’s first big data centre of excellence for the whole agri-food value chain, from farm through to consumer. Building on this, Professor Dobermann will present information on a new social enterprise currently being established in Africa, iSDA, which will focus on innovative solutions for decision agriculture.

Read More

Taking agronomy to scale with dry soil-plant spectroscopy

Abstract

Using the simplicity of light, advances in dry spectroscopy and new sensors are providing for rapid and low cost, non-destructive analysis of soil and plant materials. Mid-infrared spectroscopy provides a fingerprint of a soil’s mineralogical and organic composition, which relates to most soil functional properties, and also measures plant nitrogen and many other plant attributes. Portable x-ray fluorescence spectroscopy directly measures the elemental composition of soils and plants. Together these spectral technologies allow rapid analysis of soil properties, plant nutrient status, input quality (fertilizers, manures) and crop product quality. Coupled with advances in geo-informatics and data science, the Africa Soil Information Service is deploying soil-plant spectroscopy to provide digital mapping of soil properties at national and sub-national scales, and soil and plant testing support for large-number, multi-location agronomic trials. These advances provide an unprecedented opportunity for cost-efficient, evidenced-based, precision targeting and validation of fertilizer and other soil management recommendations.

Read More

How Big Data can improve the process of decision-making?

Abstract

Decision making is the process of making choices by identifying a decision from different choices, collecting information and evaluating alternative resolutions. The information collected is a key success factor that impacts the nature and quality of the decision. Indeed, the company’s decision makers are starting to realize the importance of using more data to support decision-making for their strategies. Therefore, it is crucial to use solutions that accurately extract useful information from the data and transform it into knowledge that will be used to make relevant decisions. It is also important to consider the revolution that occurs when data changes in volume, acquisition speed, and heterogeneity. But does Big Data mean big knowledge? How is knowledge managed with Big Data? How conventional statistical and data mining approaches handle Big Data to enable decision makers to build knowledge from Big Data?

This conference aims to discuss the concept of Big Data and its implications in the decision-making process. We will also introduce the new technologies and architectures required to store and analyze these Big Data through advanced data analytics.

Read More

Implementation of Big Data for smallholder moroccan farmer

Abstract

Precision agriculture (PA) and the use of big data include new technologies that combine instruments and information technologies to optimise inputs and improve crop productivity. These technologies are being introduced in farming practices in developed countries and have shown some benefits to the farmers. Generally, farmers in developing countries have larger farms and more resources to implement the PA technologies in their farms compared to farmers in developing countries. Furthermore, the education level to manage the hi-tech tools may be a handicap in developing countries. Can a relatively small farmer in a developing country such as Morocco where the average size of the farm is approximately 1.6 ha take advantage of the new PA technologies? The presentation will give some examples of challenges affecting productivity in Morocco and how PA could be implemented. PA technology examples include tools to manage water, soil fertility, weeds and diseases that may affect yield and quality of crops.

Read More

Hands-on Agronomy – the key to ensuring “big data” adds value to farmers

Abstract

In the rapidly advancing world of agricultural technology, big data can provide farmers with an unprecedented amount of data about their farming operations allowing them to increase yields and improve efficiency. Crop Nutrition Laboratory Services (CropNuts’), which started in 1998 as an independent soil testing laboratory, has been working at the forefront of this new wave in technology as it expands in Africa, supporting farmers to adopt precision agriculture practices. Increasingly labelled as key to transforming Africa’s agriculture, big data undoubtably has huge potential to improve agricultural efficiency in Africa if targeted effectively – from helping regions and countries to develop strategies that support their farmers in a changing climate to supporting decision making by individual farmers. This talk discusses the latter – highlighting both the opportunities and challenges for farmers using big data in the African context based on Cropnuts’ experience. It focuses on the most important lesson that CropNuts’ has learned – that the use of big data in Africa cannot replace agronomy services and, in many cases, actually increases the need for these services. Big data almost requires farmers to become data analysts themselves. In more advanced farming regions, farmers have access to real-time data, software which provides recommendations based on decades of data collection and remote support services – all of which can help them interpret data and make decisions independently.

This type of support does not exist for much of Africa. CropNuts’ is working with farmers to provide these services, helping farmers improve their own data collection whilst also conducting trials to provide site-specific recommendations to support precision farming. The availability of agronomic support services is an important consideration in the discussion on how big data can transform African agriculture in a continent where a lack of agronomic support services is often cited as one of the major barriers to agricultural development. Although enormous efforts are going into creating online tools and databases to deliver data directly into farmers hands, particularly smallholder farmers, it is essential to monitor whether and how these are being used by farmers to improve farming practices.

Read More

Agri analytics for in-season nutrient management; The case of drone-based hyper-spectral sensing to address nitrogen stress in rain-fed Maize in the Midwest of United States

Abstract

Compact hyperspectral sensors compatible with UAV platforms are becoming more readily available. These sensors provide reflectance in narrow spectral bands while covering a wide range of the electromagnetic spectrum (typically in the visible and near infrared range). This combination has the potential to identify plants with different stress levels or nutrient deficiencies. However, because of the narrow spectral bands, wide spectral range and pixel size, hyperspectral imagery can benefit greatly from data mining and machine learning techniques to leverage its power. In this study, rain-fed Maize was subjected to six different levels of nitrogen stress, and hyperspectral images were collected around sidedress time (growth stage V6) using a line scanner sensor (Resonon PIKA II) on board a hexacopter UAV (DJI Matrix 600 Pro) platform. At the same time plant tissue N were measured. A framework for the steps to extract actionable insights from the hyperspectral data will be discussed. It included pure pixel segmentation, spectral extraction and preprocessing, and six machine learning algorithms comparison. These algorithms included logistic regression (LR), support vector machine (SVM), random forest (RF), boosting, bagging, and artificial neural network (ANN). A confusion matrix based on a validation set (unseen by the models) was used to determine the performance accuracy of these classifiers and to illustrate the type and magnitude of errors for each classification method. Simulated broadband normalized difference vegetation index (NDVI) was also derived from the hyperspectral data and compared to machine learning algorithms for the prediction N stress class. Results were used to make variable rate sidedress nitrogen fertilizer recommendations.

Read More

Latest Fertility management practices – Fertigation monitoring and water managmeent

Abstract

Nowadays, farmers are facing serious issues: an increasing population, an increase in the production costs, and a decrease in many production activities. We can say each day the demand for fresh products increases while the supply decreases.

Within this framework, it becomes necessary to find better tools and methods to optimize agricultural production.
The crop fertigation monitoring service seeks to optimize resources such as water and fertilizers by using them efficiently in an environmentally-sustainable way. Through high-tech analyses and top-quality interpretation, this monitoring serves as an important decision-making tool.
The above technique plus the addition of water management, such as infiltration rate and soil retention humidity curves give the farmers the information needed to adjust the irrigation to the crop requirements, as well as the amount of fertilizers.

Through specific cases in high value crops such as avocado and blueberries, we will present and discuss how these techniques can help us understand the fertility requirements of each crop and in each cropping stage.

An additional point which would give us more precise information would be the introduction of dendrometers and soil humidity probes. These allows us to combine the soil humidity plus the plan growth and get a global understanding. Examples for the Avocados under drip by drip irrigation will be also discussed during the presentation

Read More

UAV-LiDAR for precision agriculture

Abstract

Geospatial technologies offer large opportunities to the agricultural industry. Unmanned aerial vehicles (UAVs or drones) have revolutionized agriculture from planting to harvesting, making possible to put aerial data to work in new ways, ranging from detecting crop damage to analyzing stand counts. Another technology is Light Detection And Ranging (LiDAR) that it is one of the most innovative and accurate tools providing high-resolution 3D Big Data. LiDAR is used extensively for various reasons in the agricultural sector. It’s a reliable method for monitoring and predicting the growth and yield of a crop. In our presentation, we will present these two technologies, their potentials, challenges and limits for their integration in precision agriculture.

Read More

Smart Agriculture and Internet of Things – Challenges and Opportunities

Abstract

The rapid growing of world population requires producing more agriculture products (food, medical plants, fuel, etc.) with less natural resources. Furthermore, the environment concerns as well as the new consumers’ lifestyle should be taken into consideration to ensure a sustainable agriculture. A recent concept named ‘Smart Agriculture’ (or Smart Farming) is considered as a major approach for developing sustainable agriculture. Smart Agriculture involved advanced Geospatial and Information Communication Technologies (such as GNSS, Internet of Things (IoT), Artificial Intelligence, Big Data). In particular, IoT technologies are based on smart devices that are connected through wireless network infrastructure. They allow collecting a vast amount of georeferenced environmental and crop data.

These data are then analyzed to help the farmer making efficient and accurate decision to ensure high quality crop. In this presentation, we will highlight the importance of IoT technologies in smart farming and their application in various aspects of agriculture. In addition, we will show how this connected smart objects interact together to provide valuable information for intelligent production and efficient decision making enhanced by context awareness and trigged by real-time events. Furthermore, the use of Artificial Intelligence to perform predictive analytics to track and predict environmental impact on crop yield is illustrated. Finally, various challenges that arise from the use of IoT in agriculture are discussed (Interoperability, Networking, Security).

Read More

Use of Big Data to develop rapidly and timely portfolio of crop varieties

Abstract

“Global food demand is projected to double by 2050 and this is a major and unprecedented challenge for agriculture today, in a time of rapid changing global conditions. Mathematical approaches have played a major role in advancing crop improvement since the 1920s, prior to the emergence of molecular approaches. There is today a considerable potential for using mathematical approaches in combination with Big Data and farmers’ insights to accelerate crop improvement. There is a potential to identify rapidly useful traits in genetic resources and to incorporate these identified traits into crop improvement, within a shorter period of time and with less costs. Useful traits include among others, tolerance to heat, drought, salinity and resistance to major pests and crop diseases.

Estimates show that new varieties can be resistant to diseases for an average of five years, while it generally takes 8-11 years to breed new varieties. Such transboundary diseases are causing significant losses to farmers and their outbreaks and upsurges exacerbated by climate change is threatening livelihoods.

There is a need to close the gap and to raise the potential of crop yields. In part it depends on the correct combination of specific alleles or genes at the 50–60,000 genetic loci, but mostly by the speeding up the process by which these traits are identified and incorporated into crop varieties.

Previous simulations on crop improvement show that the benefits of using improved crop varieties can be Substantial provided that the traits of interest are identified early. This paper will demonstrate also in practice how mathematical approaches in combination of Big Data have contributed in the mobilization of genetic resources in crop improvement, and how they will advance global collaborative research efforts to address current and future challenges of developing rapidly and timely portfolio of crop varieties involving farming community.”

Read More

Introduction to digital soil mapping techniques

Abstract

Soil and land surface mapping is traditionally carried out using field work and ground sampling at spot locations (e.g. elevation measurements, in-situ spectral measurements, physical sample retrievals for grain size and chemistry). The field data is then projected onto maps using spatial interpolation methods, such as linear interpolation, and more sophisticated statistical model-based interpolations, such as kriging. With the advances of Geographical Information System (GIS), the workflow of spatial mapping becomes easier and faster, providing a versatile platform for data integration that can complement soil and land surface feature recognition. There are new types of input datasets that are readily available for land surface mapping, including shallow surface geophysical methods (e.g. magnetic susceptibility, electronic resistivity) and remote sensing approaches (e.g. radar, LiDAR, multi- and hyperspectral imaging) that can be used to compliment exiting methods. Optical remote sensing can improve spectral properties through surface reflectance. Based on the surface reflectance image data can be classified using supervised and unsupervised approaches to provide spatial map for decision mapping and exploration. However, these maps cover only the surface, and they have limited implications to the subsurface properties.

Indirectly spectral data can be linked to subsurface properties through image classification and regression with sufficient ground calibration data. In particular, the hyperspectral imaging has a great use because it has many narrow spectral channels that are very sensitive to vegetation characteristics (e.g. both physical and chemical components), exposed soil fraction, and silicate and ferrous mineralogy. This presentation will explore the use of airborne, high-resolution hyperspectral imaging for soil and land surface mapping through multiple New Zealand examples. These examples provide an overview of the potential in hyperspectral imaging for environmental mapping, agriculture and forest mapping, and geological exploration.

Read More

Big data and digital soil mapping responding to the contemporary environmental issues of our time.

Abstract

This talk is concerned with practicable methodologies for delivering comprehensive spatial soil information to end-users. There is a need for relevant spatial soil information to complement objective decision-making for addressing current problems associated with soil degradation; for modelling, monitoring and measurement of particular soil services; and for the general management of soil resources. These are real-world situations, which operate at spatial scales ranging from field to global scales. As such, comprehensive spatial soil information is tailored to meet the spatial scale specifications of the end user, and is of a nature that fully characterises the whole-soil profile with associated prediction uncertainties, and where possible, both the predictions and uncertainties have been independently validated. ‘Practicable’ is an idealistic pursuit but nonetheless necessary because of a need to equip land-holders, private-sector and non-governmental stakeholders and, governmental departments including soil mapping agencies with the necessary tools to ensure wide application of the methodologies to match the demand for relevant spatial soil information. Practicable methodologies are general and computationally efficient; can be applied to a wide range of soil attributes; can handle variable qualities of data; and are effective when working with big data.

I will give a wide-ranging talk of some of the capabilities of creating comprehensive spatial soil information systems. I will first talk on methods used for the 3-D prediction of soil properties with associated quantified uncertainties. Then I will provide some examples where such methods have been used in Australia mainly, but from other places around the world too. My talk will advance on to describing some applications of using comprehensive spatial soil information systems including: digital soil assessment, in particular for agricultural enterprise suitability; and for generating user-specified soil information products in the context of farm-scale soil mapping from national scale models. Finally I will describe some new and exciting directions for delivering even better soils information. Overall my presentation will highlight how big data, data analytics and advances in statistical modeling have revolutionised soil mapping, to which has given us the enhanced ability to determine the status of soils, detect emerging environmental issues, and monitoring the condition of soils through time.

Read More

Nitrogen management: site-specific strategies to balance agronomic and environmental goals

Abstract

Fertilizer management is challenged by landscape complexity and climate variability. Yield goals often vary greatly within and across fields, and so a one-size-fits-all approach to nitrogen fertilizer management is inefficient. Site-specific fertilization strategies are needed to match nitrogen supply with variable crop demand using a three-pronged approach that involves (1) the implementation of precision agriculture technologies, (2) the utilization of spatial soil, terrain, and crop data to predict crop performance, and (3) the evaluation of site-specific practices. This presentation will feature remote sensing and other technologies to collect georeferenced crop and soil data, which can be used to assess crop performance for site-specific fertilizer management. The presentation will also highlight evaluation tools being developed to establish nitrogen use efficiency thresholds for meeting agronomic and environmental goals.

Read More

The use of big data in agro meteorology and precision farming with a special focus on remote sensing data

Abstract

“Agriculture and more precisely precision agriculture are becoming favorite areas for the exploitation of big data techniques. The sources of information used in precision agriculture are changing rapidly with the evolution of new technologies. Previously the farmer decided alone, from his experience as field worker on doses of fertilizer, periods and types of treatment against diseases or other pests of crops. Today and more and more in the future, the farmer will be advised by decision-making tools that will have processed numerous data from different origins to lead to improved management of their plot, in particular the spatial management of their plot (we do not put the same amount of fertilizer everywhere in the plot).
These data come from sensors on board tractors, drones, robots, satellite images, weather networks, crowdsourcing and, of course, the farmer’s personal knowledge.

All of this information seems to be perfectly suited to the development of new Big Data techniques in the agricultural sector. However, this type of data is very often affected by numerous errors or inaccuracies that can lead to weak or even wrong processing and therefore to a bad exploitation of big data methods.

A preliminary reflection on the quality of the input data is necessary before valorizing them correctly in Big Data systems. In addition, it also appears that some data that are a priori very interesting in big data processing are prohibited from valuation by privacy issues. All these obstacles must be considered for an efficient use of Big Data techniques.”

Read More

Weather Analytics:  Making the most of our weather

Abstract

Weather is a controlling factor in most environmental landscapes. Collecting, analyzing, and interpreting the past 15 years of weather data for agriculture and environmental applications is the focus of this presentation. The data and analytics to produce high-resolution, regional temperature maps and examples of the current and future use of these geospatial products will be presented

Read More

The use of high-throughput sequencing technologies and bioinformatics to study soil and crop microbiomes

Abstract

Beneficial soil microbes provide important ecological services to natural and agricultural ecosystems. Some groups of beneficial soil microorganisms, such as arbuscular mycorrhizal fungi, Rhizobia and endophytes have developed close symbiotic relationships with plants over time. As a result, plants are now considered as holobionts, which are inseparable from their microbial partners, similar to animals and humans. Our ability to study complex soil and crop microbiomes has been limited by technological hurdles, but high-throughput sequencing technologies and bioinformatics, now allow us to identify critical differences in microbial composition and activity between productive and unproductive soils and tracing beneficial and pathogenic microorganisms. These high-throughput technologies produce billions of DNA or RNA sequence data in a single run, leading to massive datasets that can be used in precision agriculture including the improvement of plant nutrient uptake efficiency through the management of soil microbiomes by enhancing the contribution of beneficial soil microorganisms and reducing the activity of root pathogens. This presentation will provide an overview of the applications of high-throughput sequencing methods, bioinformatics pipelines and biostatistics to study soil and crop microbiomes in agrosystems.

Read More

IoT for precision agriculture – trend and challenges

Abstract

WSN is an emergent and multidisciplinary science, a very active and competitive research area. WSN was first motivated by military applications such as battlefield surveillance, transportation monitoring, and sensing of nuclear, biological and chemical agents. It started with the smart dust project where the main goal was to implement a complete sensor and communication system integrated into a cubic millimeter package. WSN has unprecedented levels of integration and cross-layering of hardware, software and sensor (NEMS/MEMS). WSN is the foundation of Internet and Web of Things ‘IWoT’. Thanks to 6LoWPAN, RPL and CoAP new advances were achieved in communication protocols, which enable WSN nodes to connect P2P through internet. Nowadays, low power long-range wireless access mediums are available such as Sigfox, LoRa, WMBUS and NB-IoT. These new technologies come with a large panel of software frameworks and infrastructure platforms (e.g., edge, fog and cloud frameworks) to meet different application requirements. Moreover, long-range wireless access mediums increase significantly the reliability of the wireless sensor network. However, in spite of the tremendous advances many challenges need to be overcome and many questions are still open for real world large scale IWoT applications: smart home, smart care, smart grid, smart farming, industry 4.0 …

My presentation will focus on the global panorama IoT technology by considering the big main ICT actors such as IBM, INTEL, SAP, TI, ARM, HUAWEI, TENCENT, SAMSUNG and academic research institutions (e.g. SICS, Berkeley …).
Until now academic research institutions are the main contributors driving and developing the core technologies dedicated to IWoT. Therefore, it is important to provide a brief survey of the current state-of-the-art of the IWoT core technology by highlighting the challenges and the open research issues of basic hardware and software components such as WSN node hardware, operating system, 6LoWPAN and RPL, and CoAP. The key features of WSN node to be used for the large scale outdoor/indoor applications such as smart farming, smart care and CPS ‘Cyber Physical System’ will be described.
Finally, to illustrate the maturity of the IWoT core technology, a real-time demonstration of an IWoT platform dedicated to smart farming named iLive Platform will be shown.

Read More

Investment in new technologies by large agricultural companies and investors

Abstract

Food and agribusiness form a $5 trillion global industry with continuous growth. By 2050, caloric demand will increase by 70 percent, and crop demand for human consumption and animal feed will increase by at least 100 percent. Meeting this demand won’t be easy: 40 percent of water demand in 2030 is unlikely to be met, and more than 20 percent of arable land is already degraded. Luckily, we find ourselves in the middle of a scientific revolution and confluence of a broad array of technologies. New innovations are coming from digitizing agriculture – using both software and hardware applications for developing robots and machine learning to improve the entire supply chain. And advances in biotechnology and a better understanding of ecology are improving the food we eat and how it is grown. Seeing the opportunity, strategic and financial investors are racing to capture value from technological innovation and discontinuities in food and agriculture. Agtech is the small but growing segment of the startup and venture capital universe that’s aiming to improve or disrupt the industry. Investment in the current wave started in 2013 after the acquisition of Climate Corp by Monsanto for $1 billion. Following 2013, investment activity intensified, jumping from 76 investments totaling $309 million in 2013 to 153 investments and $666 million by 2014. Total value surged even higher in 2015 to $1.4 billion and was more than $1.5 billion in 2017. The US venture community is providing most of the money. In 2Q of 2017, US VCs committed a record $530 million to agtech startups, representing 47% of global capital invested in agtech that year. These investments are mostly a maturing VC ecosystem surrounding financings for late-stage startups creating the first Ag unicorns. This is prompting the discussion if we are in an “AgTech bubble”. Over the last four years, seed/angel agtech financings have also continued to grow. The 81 seed/angel financings recorded in 2016 were a 31% jump from 2015 and more than double the 31 angel/ seed rounds done in 2013. Part of the optimism at the seed stage investments comes from a lack of R&D spending in the large corporates. The role of corporate and strategic investors continues to evolve as the landscape changes and mergers close. Corporates will need to find innovation outside their companies. Four years after the first fundraising boom, the sector is starting to generate exits – mostly funded by corporates. The valuation and return data coming with them will attract institutional money. Investors and technology will continue to come from outside of the Ag industry, changing the industry forever and allowing us to meet the future needs of people and the planet.

Read More

Digital diffusion for targeting ecological intensification for resilient agroecosystems

Abstract

The technological advancements in agriculture have resulted in higher yields but lower ecological efficiency and nutritional value. Little innovations in later sectors have crippled our agro-ecosystems to meet the ever-growing demands. The agricultural is becoming the prominent sector for reaching out to the sustainable development goals and stability of farming livelihoods. The digitization of the agroecosystems become most essential entry point for any sustainable developmental entities whether it is plant genetics for breeding better varieties, crop diversification and intensification, efficient use of farm inputs and chemicals, agronomic practices, better economic return to restoring ecosystem services. Recent advances in Earth Observation System (EOS), Open-Access (AO), Artificial Intelligence (AI), Machine Learning (ML), Information, and Communication Technologies (ICTs) along with smartphone enabled Citizen Science (CS) making Big-Data analytics much smarter, interoperable and useful ever before. This has opened tremendous opportunity to address the gaps at multiple levels (e.g., data, yield, ecology, economy, resilience) for demand-driven precision interventions across the scale (e.g., space, time and package). The digital diffusion in agricultural production will help enabling infusion of ecological functions to achieve desired return (production follows functions). Ongoing efforts in big-data analytics help quantify production dynamics, trajectories and drivers under changing climate, demography and degradation process to target site specific developmental interventions and scaling the proven technologies, such as name few; intensification of food legume in rice fallows, adoption of conservation agriculture, quantification of yield gaps, land and water productivity potential and transboundary cooperation. The overarching goal is to build an integrated production system by leveraging technological diffusion with sound ecological foundation to design an ‘inclusive agroecosystems’ that gives power to farmer/primary producers to policy makers for green economy in the dry areas.

Read More

Using environmental, phenotypic and molecular data for an efficient use of genebank genetic variation

Abstract

Genebanks worldwide ensure the long-term conservation of genetic resources to be readily available for use by breeders, researchers and other users. Most of genebanks are facing major problems of size and organization. Some collections have grown so large making their conservation and use challenging. Several methodologies have been developed to respond to seed requests including random samples, core and mini-core collections which might not capture the needed diversity.

The International Center for Agricultural Research in the Dry Areas (ICARDA) in collaboration with Australian partners have developed over the past 10 years an alternative approach for better targeting of adaptive traits. The Focused Identification of Germplasm Strategy (FIGS) is a trait-based approach allowing the identification of sought traits with high probability. FIGS uses the agro-climatic and edaphic information of collection site to predict for adaptive traits of the accessions collected from the same site. This method seeks to determine and quantify relationships between collection site agro-climatic conditions and the presence of specific traits, such as disease resistance or heat tolerance using machine learning algorithms. This approach has led to the discovery of previously undiscovered genes and useful variations of known genes for resistance to serious pests and diseases. FIGS strategy is also augmented by the use of high density molecular markers for a better genebank accession’s mining and also by automatizing FIGS processes and packaging into R platform for a wider dissemination to the worldwide genebank communities.

Read More

Digitalizing data along the crop value chains: from genebank to farmers’ fields

Abstract

Innovation in plant breeding is imperative to meet the growing demand for food and feed due to global challenges such as population growth, the evolution of consumer preferences due to migration to cities and climate change. The digital revolution is a critical, and necessary, step to modernize breeding programmes. The Integrated Breeding Platform (IBP, https://www.integratedbreeding.net) has developed an integrated suite of databases, software and advanced analytical tools, known as the Breeding Management System (BMS), which allows breeders to transition into the digital era through electronic crop data storage, retrieval, analysis, sharing, interpretation and utilisation. The BMS is an example of an existing analytical pipeline, broadly adopted in a large number of developing countries, that provides breeders across a broad range of crops with access to the digital revolution. Building on the tools and services available through the IBP I will illustrate how some challenges in data management, in particular data quality, documentation, trait ontology and meta-analysis, may be overcome in an enterprise system. I will also illustrate how the BMS needs to evolve to become a modular cloud system, able to respond quickly to a moving environment and integrate applications and web services to incorporate new technologies, and manage with big data in an effective way. Last but not least I will underline the importance of integrating data to generate knowledge, when supported by appropriate analytical tools. I will use data integration along the crop value chain as a study case.

Read More

A new social enterprise providing decision-support tools for enhancing smallholder agriculture in Africa

Abstract

The Africa Soil Information Service (AfSIS) has been working over the past decade to generate digital solutions for mapping the soils and agricultural environments of sub-Saharan Africa based on field sampling, spectral soil and plant analysis, and remote sensing inference. This has resulted in a vast amount of data and a suite of novel approaches and software tools.

The primary objective of this Gates Foundation-funded program has been to generate more reliable and scalable soil information to help governments, NGOs and private sector partners make more effective strategies and policies, particularly regarding the enhancement of smallholder agriculture in sub-Saharan Africa. Recent use cases have provided critical proof-of-concept for the program.

AfSIS is now transitioning into a social enterprise in order to scale-up uptake and impact of its outputs in a financially sustainable way. Innovative Solutions for Decision Agriculture Ltd. (iSDA) will achieve this through a blended finance model combining grant funding, income generation and cross-subsidies. In this way, iSDA will be able to continue activities associated with the generation of global public goods and open source tools. In addition, the new structure and business model will enable iSDA to foster the delivery of low cost products, services and support programs that enhance the livelihoods of small-scale producers across the region and beyond. Dr Crouch will present highlights of AfSIS achievements and an overview of iSDA’s immediate priorities and long-term aspirations.

Read More

Usage of Artificial Neural Network in Sustainable Agriculture

Abstract

ANNs are inspired by the learning processes that take place in biological systems. Artificial neurons and neural networks try to imitate the working mechanisms of their biological counterparts. Learning can be perceived as an optimisation process. Biological neural learning happens by the modification of the synaptic strength. Artificial neural networks learn in the same way. The synapse strength modification rules for artificial neural networks can be derived by applying mathematical optimisation methods. Learning tasks of artificial neural networks can be reformulated as function approximation tasks. Neural networks can be considered as nonlinear function approximating tools (i.e., linear combinations of nonlinear basis functions), where the parameters of the networks should be found by applying optimisation methods. The optimisation is done with respect to the approximation error measure. In general it is enough to have a single hidden layer neural network (MLP, RBF or other) to learn the approximation of a nonlinear function. In such cases general optimisation can be applied to find the change rules for the synaptic weights.

Read More

Farming for Data; Harvesting Insights
Read More

3D models for crop monitoring in precise agriculture

Abstract

In general, efficiency of crop production is influenced by adequate field management during the growing period. This requires appropriate technical methods, but even more, knowledge about the nutrient content of plants during cultivation is necessary.
Crop monitoring can be interpreted as observing plant development during its growth. Site-specific crop parameters, such as plant height, biomass, density and coverage should be estimated for optimizing crop management.

Nowadays, various approaches address site-specific crop monitoring and management. In particular, non-contact surveys, enabled through remote sensing sensors, are in the focus of research. Biomass is an important plant parameter and approaches as to its estimation and monitoring are widely researched. Monitoring biomass and plant height with ground- or vehicle-based approaches is subject of several investigations.
Besides hyperspectral and optical sensors, the technology of Light Detection And Ranging (LIDAR) became increasingly important in a wide range of research fields, including the acquisition of vegetation parameters. LIDAR measurements allow high precision and optimal resolution. The result is a high accurate 3D point cloud from multi-temporal LIDAR surveys which is used to establish 3D crop models. A spatio temporal analyzing of these 3D models allows monitoring the crop height across the growing season. In this field, studies show that Terrestrial Laser Scanning (TLS) is suitable for capturing small objects like crops and establishing 3D crop models through multi-temporal TLS surveys.

As a first part of this contribution, we present the LIDAR technology and how it leads to crop monitoring through 3D crop models. In the second part, we present some studies dealing with crop monitoring by TLS surveys and highlight the advantage if this technology in crop monitoring compared to other techniques

Read More

Digital Soil Mapping And GIS Application For Soil Mapping in Morocco

Abstract

The aim of this contribution is to share the knowledge of new geomatics approaches used for soil mapping in Morocco that could be extended for other unmapped regions.

In fact, with the development of remote sensing and GIS based mapping techniques, Digital Soil Mapping offers more cost and time effective soil mapping methods for the rest of the unmapped parts of the Moroccan agricultural land. In fact, some DSM approaches were successfully tested in representative local environments then applied to the rest of the unmapped areas. Recently, the Ministry of Agriculture of Morocco with the collaboration of OCP, charged INRA to carry out a soil map in unstudied agricultural soils. Various zones totaling 2 million hectares were mapped at the 1/100.000 scale, by the national team within the national soil fertility map project.

The GIS processing includes three soil genesis factors: physiography, parent material and land use. Working on several pilot zones of Morocco, we generally use 30m spatial resolution Digital Elevation Model (ASTER-GDEM), equivalent scale geology/lithology data and Landsat satellite date generate physiographical units, parent material units and major land use units.
Adding to the GIS component, field and laboratory work is done with soil profile description and soil physical and chemical analysis. The terrain units are then classified through EU-landscape approach and expert knowledge.

Field check is processed using transect approach and the final soil unit map is edited at 1/100.000 scale. The output is a soil unit GIS vector layer with environmental and analytical data that implements the national soil geographic database which is very crucial for sustainable soil management in Morocco. Keywords: Digital soil mapping, GIS, ASTER imagery, DEM, Soil unit

Read More