DETECTION OF LEAF BLIGHT DISEASE IN SORGHUM USING CONVOLUTIONAL NEURAL NETWORK CLASSIFIER:A CASE STUDY

Abstract

Agriculture is the foundation of our country’s economy. Sorghum is a very popular crop and highly produced in a different part of Ethiopian regions. Sorghum can be affected by a wide variety of diseases, which can cause a serious loss of production and profitability. Leaf blight is a disease that is affecting sorghum highly where sorghum is produced. This disease is present in many humid areas where sorghum is grown like Arbaminch Gamo Gofa and other part of Ethiopian regions. Leaf blight is a dangerous disease unless it is controlled at an early stage. Traditional mechanism to identify the leaf blight of sorghum is physical observation and using chemicals. These mechanisms are inefficient, time consuming, expensive and needs an expert on the area. The model was designed based on the AlexNet pretrained model to detect leaf blight disease in sorghum. This research focused only on the detection of sorghum leaf blight which occurs on the leaf part of the sorghum. To do so, first,2000 original digital images were acquired from Gamo Gofa zone where