Evaluation of image processing technique in identifying rice blast disease in field conditions based on KNN algorithm improvement by K-means

Nowadays, rice farming is affected by various diseases that are economically significant and worthy of attention. One of these diseases is blast. Rice blast is one of the most important limiting factors in rice yield. The purpose of this study is the timely and rapid diagnosis of rice blast based on the image processing technique in field conditions. To do so, color images were prepared using image processing technique and improved KNN algorithm by K-means was used to classify the images in Lab color space to detect disease spots on rice leaves. Squared classification was based on Euclidean distance, and the Otsu method was used to perform an automatic threshold histogram of images based on shape or to reduce the gray level in binary images. Finally, to determine the efficiency of the designed algorithm, sensitivity, specificity, and overall accuracy were examined. The classification results showed that the sensitivity and specificity of the designed algorithm were 92% and 91.7%, respectively, in the determination of the number of disease spots, and 96% and 95.65% in determining the quality of disease spots. The overall accuracy of the designed algorithm was 94%. Generally, the results obtained showed that the above method has a great potential for timely diagnosis of rice blast.