An image analysis method can be used to identify nutrient deficiencies in plants by analyzing specific visual cues in plant morphology. This approach is non-invasive, rapid, and scalable, making it valuable for agricultural research and plant breeding. An input leaf image is broken up into smaller blocks using the proposed method. Second, a collection of CNNs receives each leaf pixel block. Each CNN is trained specifically for that condition to determine whether a block has any corresponding nutrient deficiency symptoms. The CNN responses are combined using a winner-take-all approach to form a unified response for each block. A multi-layer perceptron is employed to aggregate the responses from all the blocks, generating a comprehensive response for the entire leaf. and the precautions. which is to add Ca, Fe, K, Mg, and N to fill in the gaps