Agricultural UAV-based remote sensing tools to facilitate decision-making for increasing productivity in developing countries were developed and tested. Specifically, a high-quality multispectral sensor and sophisticated-yet-user-friendly data processing techniques (software) under an open-access policy were implemented. The multispectral sensor—IMAGRI-CIP—is a low-cost adaptable multi-sensor array that allows acquiring high-quality and low-SNR images from a UAV platform used to estimate vegetation indexes such as NDVI. Also, a set of software tools that included wavelet-based image alignment, image stitching, and crop classification have been implemented and made available to the remote sensing community. A validation field experiment carried out at the International Potato Center facilities (Lima, Peru) to test the developed tools is reported. A thorough comparison study with a wide-used commercial agricultural camera showed that IMAGRI-CIP provides highly correlated NDVI values (R²≥ 0.8). Additionally, an application field experiment was conducted in Kilosa, Tanzania, to test the tools in smallholder farm settings, featuring high-heterogeneous crop plots. Results showed high accuracy (> 82%) to identify 13 different crops either as mono-crop or as mixed-crops.

LEDFLEX is a micro-lidar dedicated to the measurement of vegetation fluorescence. The light source consists of 4 blue Light-Emitting Diodes (LED) to illuminate part of the canopy in order to average the spatial variability of small crops. The fluorescence emitted in response to a 5-μs width pulse is separated from the ambient light through a synchronized detection. Both the reflectance and the fluorescence of the target are acquired simultaneously in exactly the same field of view, as well as the photosynthetic active radiation and air temperature. The footprint is about 1 m² at a distance of 8 m. By increasing the number of LEDs longer ranges can be reached. The micro-lidar has been successfully applied under full sunlight conditions to establish the signature of water stress on pea (Pisum Sativum) canopy. Under well-watered conditions the diurnal cycle presents an M shape with a minimum (Fmin) at noon which is Fmin > Fo. After several days withholding watering, Fs decreases and Fmin < Fo. The same patterns were observed on mint (Menta Spicata) and sweet potatoes (Ipomoea batatas) canopies. Active fluorescence measurements with LEDFLEX produced robust fluorescence yield data as a result of the constancy of the excitation intensity and its geometry fixity. Passive methods based on Sun-Induced chlorophyll Fluorescence (SIF) that uses high-resolution spectrometers generate only flux data and are dependent on both the 3D structure of vegetation and variable irradiance conditions along the day. Parallel measurements with LEDFLEX should greatly improve the interpretation of SIF changes.

Accurate determination of plant water status is mandatory to optimize irrigation scheduling and thus maximize yield. Infrared thermography (IRT) can be used as a proxy for detecting stomatal closure as a measure of plant water stress. In this study, an open-source software (Thermal Image Processor (TIPCIP)) that includes image processing techniques such as thermal-visible image segmentation and morphological operations was developed to estimate the crop water stress index (CWSI) in potato crops. Results were compared to the CWSI derived from thermocouples where a high correlation was found ( $r_{Pearson}$ = 0.84). To evaluate the effectiveness of the software, two experiments were implemented. TIPCIP-based canopy temperature was used to estimate CWSI throughout the growing season, in a humid environment. Two treatments with different irrigation timings were established based on CWSI thresholds: 0.4 (T2) and 0.7 (T3), and compared against a control (T1, irrigated when soil moisture achieved 70% of field capacity). As a result, T2 showed no significant reduction in fresh tuber yield (34.5 ± 3.72 and 44.3 ± 2.66 t ha ${}^{-1}$ ), allowing a total water saving of 341.6 ± 63.65 and 515.7 ± 37.73 m ${}^3$ ha ${}^{-1}$ in the first and second experiment, respectively. The findings have encouraged the initiation of experiments to automate the use of the CWSI for precision irrigation using either UAVs in large settings or by adapting TIPCIP to process data from smartphone-based IRT sensors for applications in smallholder settings.

The spatial heterogeneity of soil properties has a significant impact on crop growth, making it difficult to adopt site-specific crop management practices. Traditional laboratory-based analyses are costly, and data extrapolation for mapping soil properties using high-resolution imagery becomes a computationally expensive procedure, taking days or weeks to obtain accurate results using a desktop workstation. To overcome these challenges, cloud-based solutions such as Google Earth Engine (GEE) have been used to analyze complex data with machine learning algorithms. In this study, we explored the feasibility of designing and implementing a digital soil mapping approach in the GEE platform using high-resolution reflectance imagery derived from a thermal infrared and multispectral camera Altum (MicaSense, Seattle, WA, USA). We compared a suite of multispectral-derived soil and vegetation indices with in situ measurements of physical-chemical soil properties in agricultural lands in the Peruvian Mantaro Valley. The prediction ability of several machine learning algorithms (CART, XGBoost, and Random Forest) was evaluated using R², to select the best predicted maps (R² > 0.80), for ten soil properties, including Lime, Clay, Sand, N, P, K, OM, Al, EC, and pH, using multispectral imagery and derived products such as spectral indices and a digital surface model (DSM). Our results indicate that the predictions based on spectral indices, most notably, SRI, GNDWI, NDWI, and ExG, in combination with CART and RF algorithms are superior to those based on individual spectral bands. Additionally, the DSM improves the model prediction accuracy, especially for K and Al. We demonstrate that high-resolution multispectral imagery processed in the GEE platform has the potential to develop soil properties prediction models essential in establishing adaptive soil monitoring programs for agricultural regions.