Deep Neural Network and KNN (CNN-KNN) Based Approach to Classify Mango Leaf Diseases

The ecological relevance of plants and the goods they produce is also diminished by fungal infections, impacting their economic value. The mango tree is severely afflicted by a fungus called anthracnose, most noticeable on the fruits and leaves. This paper's primary goal is to stipulate a viable system for a premature and reasonable solution for mango leaf disease detection by designing an appropriate and effective method. In recent years, digital image processing and deep neural network-based approaches have gained popularity in categorizing various mango leaf infections due to their high computational performance and identification accuracy. This paper proposed an algorithm based on deep neural network-based feature extraction and K-Nearest Neighbors Algorithm-based classification task to classify mango leaf diseases. This paper describes the possibility for CNN to extract mango leaf features on leaf images taken from the MangoleafBD dataset. The used dataset contains three types of leaf images, including healthy leaf images. The proposed method gives an accuracy level of 99.37 % at K-fold value 20. The obtained result shows that the developed model can be recommended for precise farming practices as a secondary opinion tool for mango leaf disease detection.