Differentiate Soybean Response to Off-Target Dicamba Damage Based on UAV Imagery and Machine Learning

The wide adoption of dicamba-tolerant (DT) soybean has led to numerous cases off-target dicamba damage non-DT and dicot crops. This study aimed develop a method differentiate response using unmanned-aerial-vehicle-based imagery machine learning models. Soybean lines were visually classified into three classes injury, i.e., tolerant, moderate, susceptible dicamba. A quadcopter with built-in RGB camera was used collect images field plots at height 20 m above ground level. Seven image features extracted for each plot, including canopy coverage, contrast, entropy, green leaf index, hue, saturation, triangular greenness index. Classification models based on artificial neural network (ANN) random forest (RF) algorithms developed the Significant differences feature observed among no significant across fields observed. ANN RF able precisely distinguish tolerant an overall accuracy 0.74 0.75, respectively. imagery-based classification model can be implemented in breeding program effectively phenotypic identify tolerance damage.