Deep Learning-Based Image Analysis for Detecting Nutrient Deficiencies of Tomato Plants

Tomato is one of the most popular crops worldwide. The success of a tomato crop is highly dependent on the health of the plants. Nutrient deficiency surveillance is typically conducted through visual inspections, which can be challenging for inexperienced farmers and home gardeners to work accurately. There is a growing need for a quick, reliable, and accurate nutrient deficiency identification system to address this issue. The proposed method is based on image processing and deep learning technologies that are highly effective for image classification tasks. Two models were trained using a dataset collected from tomato plants in Sri Lanka and evaluated using Mask Region-based Convolutional Neural Network (Mask R-CNN) and, You Only Look Once (YOLO) for deficiency classification, and results obtained 92% and 98% accuracy, respectively. Deficiency dispersion level expressed as a percentage using Mask R-CNN and followed by image processing techniques. Overall, this proposed system offers a convenient and accessible tool for farmers and home gardeners to monitor and maintain the health of their tomato plants, enabling them to achieve optimal yields and ensure profitable returns.