The efficiency and precision with which sickness in plants and animals may be recognised has increased. In the last five years, deep learning has outperformed standard machine learning. We aim to use artificial intelligence to learn more about illnesses and how to identify them as early as possible. This article contains details on an experiment that included image recognition, machine learning, and other techniques. In testing, the machine learning-based image recognition system functioned well in a variety of situations. Crop quality and quantity are both harmed by plant diseases. Plants may be inspected for disease by biologists and farmers, but this procedure is inaccurate and time-consuming. Leaf diseases may be classified using artificial intelligence (AI) and computer vision. In this research, two methodologies are examined and contrasted. PlantVillage pictures of apples and maize are improved using convolutional neural networks (CNN). To group these traits, a Bayesian support vector machine (SVM) classifier with high accuracy and precision is used. This will save the planet and farmers from an impending economic disaster. The approach extracts texture and colour information from dataset images using histograms of oriented gradients (HoG). Colour, texture, and depth are all combined in hybrid features. A random forest classifier is used to categorise hybrid characteristics selected using binary particle swarm optimisation. This approach selects the best output with the fewest characteristics. The assessment components are used to compare both strategies.