Information and Communication Technologies for Agriculture—Theme I: Sensors

Increasing in-farm efficiency and productivity is the main concern of farmers across the world considering the increasing demand of agricultural products and decreasing farmland. Towards that direction, the agricultural sector is being transformed in a rapid pace. Precision agriculture uses innovative technologies to increase crop yield while using lesser resources by establishing a decision management system, which uses data from the farm to control and estimate the number of resources required for a particular process with accuracy and precision. Precision agriculture is a rapidly developing area and emerging sensing technologies play an important role in it. From planting a seed to harvesting the yield, sensors can help growers with providing critical information in every stage of the production. This information can be used by growers to make key decisions to increase application efficiency and optimize inputs usage. Remote sensing systems can provide growers with large sets of data in a very short time, compared to manual data collection processes. The present chapter reviews the advances of sensing technology within farming practices and presents an overview of several categories of sensing systems that are used in agriculture.

The Sustainable Development Goals (SDGs) have been set with the vision for a better and more sustainable future for the world. Better management of earth’s resources could assist in combating poverty, inequality, environmental degradation, climate change and achieving, among others, water, and food security for our society. Soil plays a significant role in supporting the implementation of many SDGs. However, monitoring and reporting soil’s condition, although an imperative need, is yet a non-feasible, costly, and time-consuming procedure. Earth observation techniques could provide information to support many sectors of the environment and could be utilized as an alternative means for measuring soil properties. In this chapter we highlight the importance of soil in achieving the SDGs and how soil reflectance spectroscopy can support in its assessment at global, national, and local level. We provide examples from the literature where soil reflectance spectroscopy has been used in estimating soil moisture content, soil organic carbon/matter, clay content and other soil properties. We also discuss the limitations and ways forward of utilizing earth observation techniques for soil monitoring purposes.

This chapter gives a theoretical overview of various contact, proximal and remote monitoring solutions available for precision agriculture. Visual inspection of crop damage, which can be detected using these sensors, are introduced at first. Precision agriculture methodologies and sensors are reviewed with particular emphasis on variable rate fertilization. Different sensor platforms reviewed in the chapter ranged from drone images to tractor-mounted and hand-held devices, including the overview of autonomous platforms and robots in precision agriculture. After the theoretical overview a couple of use-cases are described to illustrate the most common practices of using proximal sensing sensors for precision agriculture. The use-case from Estonia demonstrates hand-held proximal sensor usage for variable rate fertilization. The use-cases from Lithuania illustrate field-scale monitoring and mapping of soil characteristics.

In pace with the fast evolving conditions in the agricultural area, this research work presents a set of easily-applicable, experimental techniques forming lightweight tools for modifying and evaluating the performance of radio modules that are based on the promising LoRa protocol, mainly in terms of radio coverage, energy consumption and packet losses. The proposed approach is exploiting the modern, low-cost, innovative card-sized computer systems. These systems, if combined with simple electronic components and conventional smart phone devices, can form an effective, flexible and portable measurement testbed, able to deliver results of satisfactory accuracy. In terms of software, this testbed is using wide-spread programming environments, both textual and visual, to form simple commands for controlling and monitoring the behavior of the radio interfaces of interest. The whole arrangement is ideal for educational purposes or for usage by not very experienced personnel, in the open field (e.g., orchards), the greenhouses or the livestock farming environment. This chapter, apart from highlighting the software and hardware details of the proposed measuring system and from presenting characteristic performance results, also reports on the feasibility of the LoRa radio modules to carry out the information needed to accomplish simple agricultural tasks (both sensing and acting ones), through designing and implementing typical examples.

Wireless Sensor Networks (WSNs) in the agricultural domain tend to increase with the implementation of Internet of Things (IoT) technologies and the adoption of 5G networks. Although several works study those subjects, only few focus specifically on animal production, a sector of the agricultural domain with specific requirements. In this work, the requirements for the implementation of WSNs in cattle weighing in the farm are gathered, a system is proposed to address those requirements, and it is evaluated using different routing strategies. The scenarios evaluated were: (i) use of the Collection Tree Protocol (CTP); use of the IPv6 Routing Protocol for Low-power and lossy networks (RPL); and (iii) use of Software-Defined Networking (SDN) technology. Although the packet delivery rate was similar among the scenarios, the last one provided the lower latency, with an improvement of 48% compared to scenario (ii) and 58% compared to scenario (i). The use of SDN resulted in a considerable improvement in latency, which is essential for this domain. The main implications of these results are discussed and their main limitations along with future works that could improve and expand on them.

Due to the rapid urbanization in many developing countries along with other major factors, such as global warming, the agriculture industries are striving to cope nowadays more than ever. This chapter gives a better insight on how the Information Communications Technologies and the development of Internet of Things applications can be used in the field of agriculture. While presenting general information regarding issues faced by these industries, the chapter focuses on analyzing the various low-cost hardware components (sensors, processing units, etc.), a simple Wireless Sensor Network synchronization method and a cloud/fog architecture suitable for environmental monitoring and agricultural applications. The effectiveness of the synchronization process is evaluated with experiments previously conducted on an olive grove. Likewise, the cloud/fog architecture is evaluated based on previous experimental runs led by other researchers. The results of both approaches are briefly presented and analyzed. The performance evaluation metrics used to outline their effectiveness are the Round Trip Time of a data packet and the packet load on every time step. A brief review of the key components of each of the aforementioned subjects is presented and possible future directions are provided.

The increasing world population directed food production towards precision agriculture in the recent past. In agronomy, there is an obvious growing interest for monitoring crop development using different spectral vegetation indices derived by different sensor devices. These sensors can offer a valuable perspective both at the field-scale and at the plant level. This paper aims to promote fusion of data derived by different sensors for agricultural applications comparing two novel sensing approaches for crop monitoring; (a) a recently developed active, multispectral, handheld proximal sensor named Plant-O-Meter, and (b) Sentinel-2 satellite, which carries a multispectral optical instrument. Both sensors follow the same basic measurement principles. Their operation is based on the estimation of the proportion of radiation that is reflected from the target, which in agricultural systems refers to plants or the soil, at different wavelengths of the spectrum of light. In this study, a maize field was monitored on several dates in 2018 growing season using both the Plant-O-Meter measurements and Sentinel-2 imagery. By utilizing appropriate formulas and spectral channels, various vegetation indices were calculated, and results were compared using linear regression analysis. The first results showed good positive correlations between the indices obtained by the two sensors which signify their joint potential.

National and International space agencies are determined to keep their fingers on the pulse of crop monitoring through Earth Observation (EO) satellites, which is typically tackled with optical imagery. In this regard, there has long been a trade-off between repetition time and spatial resolution. Another limitation of optical remotely sensed data is their typical discontinuity in time, caused by cloud cover or adverse atmospheric effects. Enduring clouds over agricultural fields can mask key stages of crop growth, leading to uncertainties in crop monitoring practices such as yield predictions. Gap-filling methods can provide a key solution for accurate crop phenology characterization. This chapter first provides a historical overview of EO missions dedicated to crop monitoring. Then, it addresses the rapidly evolving fields of gap-filling and land surface phenology (LSP) metrics calculation using a new in-house developed toolbox, DATimeS. These techniques have been put into practice for homogeneous and heterogeneous demonstration landscapes over the United States. Time series of Difference Vegetation Index (DVI) were processed from two EO data sources: high spatial resolution Sentinel-2 and, low spatial resolution MODIS data. LSP metrics such as start and end of season were calculated after gap filling processing at 1km resolution. Over the homogeneous area both S2 and MODIS are well able to capture the phenology trends of the dominant crop and LSP metrics were successfully mapped. Conversely, the MODIS dataset presented more difficulties than S2 to capture the phenology trend of winter wheat over heterogeneous landscape.

Over the years, the detection and classification of crown trees raised much interest for the scientists from the forest and environmental sciences, due to their essential role for landscape ecology and forestry management. Besides crown trees delineation and trees classification, an essential part of the trees health assessment is the extraction and estimation of vegetation indices (VI). These VI are exploiting the differences between the visible spectrum (RGB) and the near-infrared spectrum (NIR). Approximately a decade ago, detection of individual trees was focused on Lidar coupled with high-resolution ortho imagery or hyperspectral images, but with the increasing availability of unmanned aerial vehicles this gradual shifted towards Structure from Motion application. Building on the advantages of drone technology and the latest deep learning algorithms, the present study aims at assessing the combined use of ML techniques with spectral VIs derived from visible cameras mounted on drones, to be used as a proxy to characterise vegetation health of individual trees in an orchard field. To accomplish the study objectives, several image processing methods were implemented to the acquired drone data. The tree’s crown was extracted the crown trees by applying a deep learning object instance segmentation method. For mapping, the vegetation health, VI from visible spectrum (Red, Green and Blue) were used. Very good results were obtained in the case of the plum, apricot and walnut trees, mostly because these trees have the leaves oriented towards the camera and the spaces between leaves and branches are much smaller in comparison with the olives trees. Less reliable results were obtained for olive trees crown delineation because of its specific texture with small leaves and large spaces between branches. The signal received from the ground had an essential influence in the assessment of the vegetation health status by increasing the (Green Leaf Index) GLI mean values with a small fraction. Overall, the study demonstrates the real potential of drone applications and deep learning methods for spatial and temporal rapid assessment of trees vegetation heath.

The use of remote sensing in agriculture is expanding due to innovation in sensors and platforms. Uncrewed aerial vehicles (UAVs), CubeSats, and robot mounted proximal phenotyping sensors all feature in this drive. Common threads include a focus on high spatial and spectral resolution coupled with the use of machine learning methods for relating observations to crop parameters. As the best-known vegetation index, the normalized difference vegetation index (NDVI), which quantifies the difference in canopy scattering in the near-infrared and photosynthetic light absorption in the red, is at the front of this drive. Importantly, there are decades of research on the physical principals of the NDVI, relating to soil, structural and measurement geometry effects. Here, the gap between the historical research, grounded in physically based theory, and the recent field-based developments is bridged, to ask the question: What does field sensed NDVI tell us about crops? This question is answered with data from two crop sites featuring field mounted spectral reflectance sensors and a UAV-based spectroscopy system. The results show how ecosystem processes can be followed using the NDVI, but also how crop structure and soil reflectance affects data collected in wavelength space.

In agriculture, site-specific field management is based on several components including information regarding soil heterogeneity. Mobile geophysical sensors are useful tools to efficiently map the spatial distribution of physical parameters (e.g., electrical conductivity) for large areas (i.e., several hectares and more). In combination with the analysis of soil samples collected at selected points, these maps represent a database for decision-making, e.g., for programming and controlling fertilizer spreaders. In addition, multi-channel instruments not only provide data regarding lateral changes within a certain depth range, rather they also allow for reliable imaging of possible soil layers within the depth of investigation. The reliability of the geophysical parameter models (e.g., electrical conductivity model) is controlled by the conductivity and their contrasts and by the used sensor. A case study is presented in which the electromagnetic sensor DUALEM-21 is used at an agricultural field characterized by sandy soils (i.e., low electrical conductivity) including a relative homogeneous topsoil (i.e., with only minor differentiation). At our test site, the final geophysical parameter maps generated using DUALEM-21, are compared to the results obtained using a rolling electrode system (Geophilus) and ground-penetrating radar (GPR). All methods reveal similar patterns of soil heterogeneity. In addition, the conductivity-depth models resulting from kinematic surveys are in good agreement with those from conventional static electrical measurements and with the data from soil analysis of borehole samples. Thus, the results of our case study demonstrate that the DUALEM-21 can also be used successfully on sandy soils to map the lateral conductivity variation and possible layers within the root zone (i.e., the upper 1.5 m).

Recent developments in agricultural technologies have made available for use by the farmers a variety of sensors and sensing services. Remote sensing has become particularly popular especially after the release of free satellite images form several vendors across the globe. In addition, the use of unmanned aerial systems (UAS) equipped with diverse optical sensors is getting very popular for field scouting and mapping applications in agriculture since the unmanned aerial vehicles (UAV) have become cost-affordable to almost any farmer. To many farmers, the UAVs equipped with optical sensing systems seem like hi-tech toys which can offer detailed insight of in-field hotspots. However, most satellite and UAV derived observations are based on passive sensing systems and require high level data pre-processing before used in the field. Therefore, the data processing requirements work as a constraint for most farmers, while the limitations of the passive sensing systems that are affected by the weather and atmospheric conditions, make them unpractical when on-the-go farming applications, such as variable rate spraying or fertilizing, are needed. During the past decades, active proximal sensing has been increasingly used to provide information about canopy properties and take real-time decisions in a large range of crops. Numerous proximal sensing instruments have been developed and are commercially available. However, there are several limitations in the use of most of these devices, such as high complexity in the operation and data processing, high cost, poor accuracy, etc., that work as barriers in the adoption of these devices by small and medium size farms. Therefore, there is still room for new advancements in the development of new more cost effective and farmer friendly proximal sensing solutions. In this study a new low cost, active multispectral optical device named Plant-O-Meter was tested in real conditions comparing it with the well-proven GreenSeeker handheld device. The latter sensor is a widely used commercial canopy sensor well-accepted both by the farmers and the scientific community. It was selected as a reference sensor in the study as it works using the same operating principles, is relatively low cost and has similar measuring characteristics to the Plant-O-Meter. The study took place at two experimental fields cultivated with maize (Zea mays L.) using a randomized complete block design with three replications. Nitrogen (N) fertilization rate experiments were set in order to create variations in canopy development, vigor and greenness across the fields, providing the ability to compare sensors’ detectability and other performance characteristics in simulated field conditions. Thus, a wide range of sensor readings, from very low to very high, was expected. Treatments included five nitrogen (N) fertilization rates (0, 50, 100, 150 and 200 kg of N ha⁻¹) applied during sowing. Three maize hybrids were scanned for Normalized Difference Vegetation Index (NDVI) using both Plant-O-Meter and GreenSeeker sensors at V4, V6 and V8 growth stages. During full maturity, the central part of each plot was hand-harvested for grain (two middle rows 6 m long). Based on the present findings, the optimum timing for scanning using GreenSeeker or Plant-O-Meter was between V7 and V8 stage. Measuring within this growth stage window good estimation of end-of-season yield was achieved. In addition, the overall results indicated that NDVI obtained using GreenSeeker were quite similar to the NDVI measured by the Plant-O-Meter showing an almost 1:1 relationship. These results indicate that Plant-O-Meter exhibits strong potential for accurate plant canopy measurements and for real time variable rate fertilization applications in maize.