Real-Time Sorting Optimization Through Hybrid Contour-Shape Analysis in Automated Cocoon Processing Systems

Abstract

Automated sorting systems are pivotal in modern industrial applications, where the dual demands of speed and precision are critical, especially in dynamic, high-throughput environments like agricultural processing. This study introduces an innovative hybrid contourshape analysis method that integrates OpenCV-based contour processing with advanced machine learning classifiers to enhance real-time sorting of silkworm cocoons. By combining adaptive thresholding, multi-feature extraction techniques, and robust classification algorithms such as Support Vector Machines (SVM) and Random Forest (RF), the system achieves an impressive sorting accuracy of 94.2% at a rate of 85 cocoons per minute. This performance significantly surpasses traditional single-modality approaches, which often struggle with variable conditions. Computational experiments conducted across diverse scenarios validate the system’s robustness under fluctuating lighting conditions (e.g., 800–1600 lux) and rotational dynamics (up to 30 RPM), addressing longstanding challenges in high-throughput agricultural automation. The approach reduces error rates by 15% compared to conventional methods and offers a scalable framework for other industries, such as food processing and textile manufacturing. This hybrid method not only optimizes sorting efficiency but also minimizes computational overhead, making it suitable for deployment on resource-constrained devices like Raspberry Pi, paving the way for cost-effective automation solutions in agriculture