Automatic Wheat Leaf Rust Detection and Grading Diagnosis via Embedded Image Processing System

Wheat leaf rust is one of the major fungal diseases that makes severe drop in the wheat production. With the desire for rapid and accurate identification of disease, real-time access to the degree of that and timely measures, a wheat leaf diseases detection system based on embedded image recognition technology is designed. The system adopted the ARM9 processor with the embedded Linux platform as the main body, and the program is developed in the Qt integrated environment. At first, the captured clear disease picture of wheat leaf rust was transformed into G single channel gray image of RGB model. Second, based on exploration Sobel operator method, vertical edges detection were implemented on the gray image, eliminate the background of image and extracts binary feature point set of disease spot. Third, the noisy points in the point set is further filtered out by flood filling algorithm. Finally, the area ratio of disease spots and leaf is calculated to get the precise diagnosis of disease level. It has been verified that the recognition rate reaches 96.2% and accuracy rated reaches 92.3% with the method of image processing, which is approximately equal to the result of human vision. This system can be used as agricultural robot to inspect in the field, realizing the intelligentization in detection, identification, diagnosis and classification of crop disease.