Enhancing Crop Health: CNN-SVM Fusion for Sugarcane Leaf Disease Analysis

In today's farm technology world, we need to get better at spotting sick plants like sugarcane. So, people have found creative ways to utilize machine learning to help with this issue. This study is the first to find and sort out diseases that affect sugarcane leaves. It uses a team effort between convolutional neural networks (CNN) and support vector machines (SVM). The goal was to make a model to spot different types of sugarcane leaf sickness. This would help manage diseases better and increase crop yields faster. The study used a large set of sugarcane leaf photos. It included five different illnesses. Every picture was carefully cleaned up and put into the CNN-SVM system. The model was built with three layers for finding significant features, then two limiting layers to simplify it, and finally, two attachment points before the SVM classification part. Our findings showed remarkable skill in identifying diseases across all categories. The precision, recall, and F1-score for each class were as follows: Class 1 (84.13%, 90.02%, 86.97%), Class 2 (85.80%, 90.09%). More studies using big, small, and meaningful averages gave a complete look at the model's performance. The average scores for getting it right, finding everything, and combining both were 85.82%, 86.20%, and 85.97%. Each was calculated using different measures, but they all aimed to evaluate performance in the same way, so there wouldn't be any confusion or bias when comparing results from these texts with those found elsewhere." The total percentages were 86.14% for accuracy, 86.09% for finding everything and an F-score average of 86.07%, while overall micro averages showed a flat level at around.