USING DEEP LEARNING FOR IMAGE-BASED CROP DISEASE DETECTION IN ABUJA, NIGERIA
Author(s): Anjorin, Toba Samuel, Ogbonna, Onyedikachi Benjamin, Apeh, Ayo Igoche
Volume/Issue: Volume 3 , Issue 1 (2023)
Image-based plant leaves diseases detector for five Nigerian crops were developed using deep learning approach. A locally collected dataset of 268 images of diseased and healthy plant leaves collected under controlled conditions were used. A deep convolutional neural network was trained to identify the five crop diseases namely Sorghum anthracnose, Cowpea Cercospora leaf spot, Maize phosphorus deficiency, Rice brown leaf spot and Sunflower leaf blight. The trained model achieved a top accuracy of 94.03% i.e. 940 out of 1000 images on a held-out test set, demonstrating the feasibility of this approach. Overall, the approach of training deep learning models on increasingly large and publicly available image datasets presents a clear path toward smartphone-assisted crop disease detection and management in Nigeria and on a massive global scale.