Abstract

A low-cost detection system based on LED is designed, utilizing an attention mechanism to identify 10 different rotation speeds of multi-opening objects, achieving an average identification accuracy of 90.36% across 14 types of objects.