AN INTEGRATED BEST SOLUTIONS FOR OPTIMIZED PEST CONTROL AND SOIL HEALTH IN SUGARCANE CROPS

Abstract

In India, pest management is of great importance in order to have healthy sugarcane crops and to maximize production. Image processing and object detection models are used in this work to engage in proper sugarcane pest management. Through integrating YOLOv8m into detecting objects correctly, RandAugment to perform robust image enhancement, and ESRGAN (Enhanced Super-Resolution Generative Adversarial Networks) to produce more resolution crop images, the present study can produce an incredibly efficient pest control and early prediction system of sugarcane plant. Current soil crawling robots with sophisticated sensors can offer information on soil characteristics and preliminary pest detection, reducing the amount of crops destroyed and wasteful application of resources. The proposed system will help to predict and intervene early thus reducing losses in crops and increasing the total output of the farms. In this case, high-resolution pictures captured by the drone were taken as the input, in particular with sugarcane fields; YOLOv8m to identify pests, including stem borers, aphids as well as top shoot borers and whiteflies; apply RandAugment to teach the pictures that assist in simplifying the model and detect pests under different conditions and forms; and use ESRGAN technique to sharpen the drone images. The existing higher-resolution images may give finer information, making it easier to detect small pests or early disease symptoms by applying both traditional and proposed algorithms. The research provides a comparison analysis of traditional algorithms with YOLOv8m and similarly, a comparison of various YOLO versions with the YOLOv8m algorithm. The findings show that these integrated systems can produce great amounts of improvement in crop yield and soil health with a rate of 93%. GIS and IoT approaches enable proper allocation of resources and reduce environmental impact. Improvements in ML models have increased the pest prediction speed and speed as a recommendation in favor of sustainable pest control. The combination of AI/ML, and robotics will offer the opportunities to have sustainable development of sugarcane plants, increase the harvest, improve soil quality, and reduce the ecological footprint.