A Plant-By-Plant Level Remote Sensing Classification Method for Cotton Root Rot Based on Uav Platform

Cotton root rot (CRR), caused by fungus Phymatotrichpsis omnivore, is one of the most destructive cotton diseases in Texas. Once the plant is infected by CRR, it will be hardly cured. However, a fungicide named Topguard Terra was proved efficient to protect cotton from being infected by CRR. Previous researches indicated that the CRR will reoccur at the same region as past years. Therefore, knowing CRR-infested area is helpful to prevent CRR from appearing. The CRR-infested plants can be detected by using aerial remote sensing (RS). As unmanned aerial vehicle (UAV) was introduced to remote sensing research field, the spatial resolution of imagery data was increased significantly，which makes higher precision CRR classification possible. An algorithm based on concept of Superpixel is developed to detect and delineate CRR-infested area in cotton field at plant-by-plant (PBP) level precision. The algorithm can detect healthy or CRR-infested single individual plant automatically based on multispectral or RGB images. 5-band multispectral data were collected using UAV to test the algorithm.