Study on Drone Monitoring Project Based on Cluster Analysis and Volov5

This paper presents a five-year drone monitoring project aimed at addressing the threat to global biodiversity and ecological balance posed by the illegal trade in wildlife. The project selects the Wyoming state government as the target customer, develops a cluster analysis-based drone monitoring deployment and combines it with a YOLOv5-based identification system, and synthesizes objective elements such as customers, resources, and markets. The k-means and DBSCAN clustering algorithms were used to optimize the deployment of UAVs and solve the ecoregion intersection and spherical region constraint problems. Meanwhile, the optimal network architecture was selected by combining the VOLOv5 recognition system, PyTorch framework, and the image recognition was optimized by combining the CSP structure, Focus module, and neck-layer FPN+PAN structure to improve the detection efficiency. The model has the advantage of being comprehensive, data-driven, and multi-validated, while pointing out the disadvantages of simplifying assumptions and providing a detailed mechanistic analysis of system scheduling to make the model generalizable.