Research on Graded Diagnosis of Lettuce Water-Nitrogen Stress and Pest Prevention Based on Deep Learning

Abstract

Abstract Lettuce is extremely sensitive to the demand for water and nitrogen throughout its growth phase. These elements are crucial for plant photosynthesis and resistance against diseases and pests. When lettuce experiences water or nitrogen stress, it leads to a decline in quality, with the leaves turning yellow, severely affecting its quality, yield, and market value. Therefore, it becomes particularly important to swiftly and accurately identify the water and nitrogen stress status of lettuce and to carry out precise graded diagnostics. This paper will explore in detail the methods of recognition and graded diagnostics for the water-nitrogen stress status of lettuce. Deep learning models were constructed separately through methods based on phenotypic characteristics, image features, and target detection, integrating attention mechanisms into the models. The DNN network model established using the Feature Phenotype Dataset (FPD) can efficiently and rapidly accomplish the recognition and graded diagnostics of lettuce water-nitrogen stress status, with an accuracy of 0.8302, providing a feasible basis for monitoring and diagnosing water-nitrogen stress status using non-image data. The ResNet50Evo-SE model, which uses image data combined with the Channel-wise Attention Mechanism (SE), achieves an accuracy of 0.9897 for the recognition and graded diagnostics of water-nitrogen stress status. The YoloV8-CBAM target detection model, which uses image data combined with the Convolutional Block Attention Module (CBAM), can make precise judgements and detections of the characteristics of water-nitrogen stress status, with an overall accuracy of 0.9653, recall of 0.9820, and mAP (mean average precision) of 0.9869 and 0.7711, respectively. The YoloV8-CBAM model can also effectively identify and detect lettuce leaf miner pest damage, with an early-stage pest detection accuracy of 0.9443, and recall of 0.9531. These methods provide reliable technical support for optimizing planting management and advancing smart agriculture, contributing to the efficient and high-quality development of precision agriculture.