Lav R. Khot

Field-based crop phenotyping

High-resolution multispectral aerial imaging was used to estimate winter wheat growth parameters.Visual ratings of emergence and spring stand were compared with data extracted from aerial images.A high correlation (r=0.86) between the ground-truth and aerial image data was observed.UAV-based sensing can be an alternative to standard methods for rapid field-based crop phenotyping. The physical growing environment of winter wheat can critically be affected by micro-climatic and seasonal changes in a given agroclimatic zone. Therefore, winter wheat breeding efforts across the globe focus heavily on emergence and winter survival, as these traits must first be accomplished before yield potential can be evaluated. In this study, multispectral imaging using unmanned aerial vehicle was investigated for evaluation of seedling emergence and spring stand (an estimate of winter survival) of three winter wheat market classes in Washington State. The studied market classes were soft white club, hard red, and soft white winter wheat varieties. Strong correlation between the ground-truth and aerial image-based emergence (Pearson correlation coefficient, r=0.87) and spring stand (r=0.86) estimates was established. Overall, aerial sensing technique can be a useful tool to evaluate emergence and spring stand phenotypic traits. Also, the image database can serve as a virtual record during winter wheat variety development and may be used to evaluate the variety performance over the study years.

UAS imaging-based decision tools for arid winter wheat and irrigated potato production management

ABSTRACT Small unmanned aerial systems (UAS) are gaining global attention for rapid image-based decision making in agricultural production. In this study, the aim was to evaluate UAS-based imagery for rapid assessment of wheat winter survival and spring stand in winter wheat production and crop necrosis in potato production. Both are critical aspects of field (arid) and row (irrigated) crop farming practices. Aerial images from 97 hard and 352 soft single nucleotide polymorphism winter wheat plots, and 32 potato field plots (with 1 and 2 years of green manure applications) were acquired using a multi-band imaging sensor integrated with UAS. The UAS-based imagery was useful in evaluating winter wheat plant winter survival and spring stand, with Pearson correlation coefficient (r) in the range 0.60–0.82 between imagery and ground reference data. Similarly, the image-based potato field necrosis assessment showed a strong relationship with ground reference data (r = 0.93 and 0.88 for 1 and 2 years of green manure application, respectively). Overall, UAS imagery provided quantifiable, timely, and unbiased field data with high spatial resolution (about 2.3 cm/pixel for images acquired at 100 m altitude) that can aid in field and row crop production decision making.

High Resolution Multispectral and Thermal Remote Sensing-Based Water Stress Assessment in Subsurface Irrigated Grapevines

Precision irrigation management is based on the accuracy and feasibility of sensor data assessing the plant water status. Multispectral and thermal infrared images acquired from an unmanned aerial vehicle (UAV) were analyzed to evaluate the applicability of the data in the assessment of variants of subsurface irrigation configurations. The study was carried out in a Cabernet Sauvignon orchard located near Benton City, Washington. Plants were subsurface irrigated at a 30, 60, and 90 cm depth, with 15%, 30%, and 60% irrigation of the standard irrigation level as determined by the grower in commercial production management. Half of the plots were irrigated using pulse irrigation and the other half using continuous irrigation techniques. The treatments were compared to the control plots that received standard surface irrigation at a continuous rate. The results showed differences in fruit yield when the control was compared to deficit irrigated treatments (15%, 30%, 60% of standard irrigation), while no differences were found for comparisons of the techniques (pulse, continuous) or depths of irrigation (30, 60, 90 cm). Leaf stomatal conductance of control and 60% irrigation treatments were statistically different compared to treatments receiving 30% and 15% irrigation. The normalized difference vegetation index (NDVI), green normalized difference vegetation index (GNDVI), and canopy temperature were correlated to fruit yield and leaf stomatal conductance. Significant correlations (p < 0.01) were observed between NDVI, GNDVI, and canopy temperature with fruit yield (Pearson’s correlation coefficient, r = 0.68, 0.73, and −0.83, respectively), and with leaf stomatal conductance (r = 0.56, 0.65, and −0.63, respectively) at 44 days before harvest. This study demonstrates the potential of using low-altitude multispectral and thermal imagery data in the assessment of irrigation techniques and relative degree of plant water stress. In addition, results provide a feasibility analysis of our hypothesis that thermal infrared images can be used as a rapid tool to estimate leaf stomatal conductance, indicative of the spatial variation in the vineyard. This is critically important, as such data will provide a near real-time crop stress assessment for better irrigation management/scheduling in wine grape production.

Efficacy of unmanned helicopter in rainwater removal from cherry canopies

Developed was an in-field sensing system to monitor cherry canopy micro-climate.Unmanned helicopter was evaluated for canopy rainwater removal.Effect of flight altitude and payloads on rainwater removal was quantified. Rain-induced fruit cracking causes significant economic loss for fresh market sweet cherry growers annually. To prevent cherry cracking, timely removal of rainwater from fruit is the key. This study evaluated the efficacy of an unmanned middle-size helicopter to remove rainwater from Y-trellised cherry canopies. Helicopter downwash in hover at four altitudes, with and without a payload, was quantified with six anemometers deployed in tree canopies. Results showed that payload and altitude significantly affected hover downwash, which was greater at higher altitude of 7.6m above ground level (AGL) than lower altitude of 4.9m AGL with payload. In the absence of payload, hover downwash peaked at the altitude of 6.1m AGL. In the efficacy study, 5.0-mm rainwater was applied to cherry canopies by a rainfall simulation system, followed by the helicopter flying over canopies at three altitudes (4.9, 5.5 and 6.1m AGL), two travel speeds (1.3 and 2.7ms-1) and with or without payload. Rainwater removal at bottom (1.1m), middle (1.9m) and top (2.7m) of the canopies was calculated based on the change of leaf wetness of target canopies in 10min after rain. Overall, helicopter with payload flying 2.7ms-1 at 6.1m AGL removed significantly more rainwater (96.3%) from top section of canopies than groups without treatment (71.2%) and compared to other payload and travel speed conditions. Results also confirmed that the unmanned helicopter could provide sufficient downwash to remove rainwater effectively from bottom and middle canopy sections.

Technical Note: Spray Pattern Investigation of an Axial-Fan Airblast Precision Sprayer Using a Modified Vertical Patternator

Evaluation of the spray patterns is essential for making adjustments to an agricultural sprayer that would result in less chemical usage, less spray run-off onto the ground, and increased spray targeting accuracy. Therefore, in this study, a vertical patternator was fabricated to evaluate an axial-fan airblast sprayer. The airblast sprayer was retrofitted with variable rate nozzles and adjustable air-assist flow control, for citrus tree-specific precision spraying. Tests involved evaluating spray patterns at different nozzle flow rates and air-assist settings. Also, the air-assist measurements between the pair of nozzles were measured while the sprayer was stationary (no spray).

NAVIGATIONAL CONTEXT RECOGNITION FOR AN AUTONOMOUS ROBOT IN A SIMULATED TREE PLANTATION

A sensor fusion technique was developed for estimating the navigational posture of a skid-steered mobile robot in a simulated tree plantation nursery. Real-time kinematic GPS (RTK-GPS) and dynamic measurement unit (DMU) sensors were used to determine the position and orientation of the robot, while a laser range finder was used to locate the tree positions within a selected range. The RTK-GPS error was modeled by a second-order autoregressive model, and error states were incorporated into extended Kalman filter (EKF) design. Through EKF filtering, the mean and standard deviation of error in the easting direction decreased from 4.05 to 2.21 cm and from 8.27 to 1.89 cm, respectively, while in the northing direction, they decreased from 4.64 to 1.81 cm and from 11 to 2.16 cm, respectively. The geo-referenced tree positions along the navigational paths were also recovered by using a K-means clustering algorithm, achieving an average error of tree position estimates of 4.4 cm. The developed sensor fusion algorithm was proven to be capable of recognizing and reconstructing the navigational environment of a simulated tree plantation, which offers a great potential in improving the applicability of an autonomous robot to operate in nursery tree plantations for operations such as intra-row mechanical weeding.

Comparison of two multiband cameras for use on small UAVs in agriculture

This study evaluates the applicability of two multiband cameras used with a small unmanned aerial vehicle (UAV) for stress detection in citrus orchards. The aerial images were acquired using both cameras at UAV flying altitudes of 30, 60, and 90 m were processed to extract histogram distributions of green normalized difference vegetative index as feature datasets. Support vector machine based classification results revealed that the high resolution camera with near infrared (670–750 nm) and green bands was better in detecting healthy and unhealthy citrus trees. The highest average overall classification accuracy of 91±7% (mean ± standard deviation) was obtained using feature datasets of high resolution camera images acquired at an altitude of 60 m.

Effect of Atmospheric Conditions on Coverage of Fogger Applications in a Desert Surface Boundary Layer

Near-ground aerosol fogs were applied in the Chihuahua Desert of New Mexico, which has widely spaced, low shrub vegetation. Near-ground fog dispersion was measured remotely with a light detection and ranging (lidar) system. Local atmospheric turbulence and stability were continuously measured with 3-axis sonic anemometers during aerosol treatments. Lidar-measured plume area coverage and spread were related to the simultaneous local-scale weather, including both convective boundary layers (CBL) and stable boundary layers (SBL). A modified bulk stability ratio (SRm) was used to characterize the stability conditions near the ground. Time averages appropriate to the SBL were determined using the multidimensional decomposition technique and matched to the short spray time periods in the CBL. The widest, most effective, near-ground coverage was obtained from insect fogger applications conducted during relatively high wind speeds: U > 1 m s-1 in stable conditions, and U > 3 m s-1 in unstable conditions. In general, spraying during SBLs was more efficient than during CBLs, with less material wasted and better consistency of coverage in the target zone nearest the ground. There was no significant difference in spray coverage or plume dispersion between the handheld thermal fogger and the ultra-low volume (cold fogger) applicator used.

Solar and Storage Degradations of Oil- and Water-Soluble Fluorescent Dyes

The assessment of spray deposit efficiency is an important aspect in pesticide application technology research. Since fluorescent tracer dyes serve as useful markers for evaluating spray deposits in many spray application investigations, it is necessary to study the stability and degradation of the fluorescent deposits during storage and exposure to solar radiation before using them under field conditions. Therefore, this study was carried out to evaluate the degradation characteristics of spray deposits prepared with two fluorescent dyes (oil-soluble Yellow 131SC®, and water-soluble Pyranine 10G®), due to the solar radiation and duration of sample storage.

Neural-Network-Based Classification of Meat: Evaluation of Techniques to Overcome Small Dataset Problems

One of the objectives of our multidisciplinary research group is to develop sensors for the detection of Salmonella contamination in beef. Similar to most biological studies, beef contamination classification studies using artificial neural networks (ANNs) are restricted to small datasets. This study evaluates selected techniques of data domain expansion and synthetic sample generation on small datasets associated with meat contamination. Mega-trend-diffusion (MTD) and functional virtual population (FVP) techniques for data domain expansion and synthetic sample generation were assessed on the small datasets. The datasets used were obtained from a thin-film (TF) module electronic nose system in response to the headspace of control and Salmonella-inoculated packaged meat samples. Back-propagation neural networks (BPNNs) were used to determine classification accuracies of the synthetically expanded datasets. For aged beef datasets, the maximum mean of average overall classification accuracies provided by FVP technique was 90%. The maximum mean of average overall classification accuracies obtained by FVP technique was about 81% for fresh beef datasets. MTD technique also provided similar accuracies (in the lower 80s). Both techniques were found useful for expanding the domain range of the small dataset in order to test and evaluate BPNN-based classification models.