Research on items sorting robot based on SSD target detection

Abstract

To achieve the automatic classification of items and improve the sorting efficiency of the sorting robot. A design method of a robot based on color classification is proposed. The system uses an all-around Mcnamu wheeled mobile platform equipped with a Raspberry Pi 4B, an STM32 MCU, a USB monocular camera, and a 6-DOF robotic arm. The SSD target detection model is used to identify sorted items, and color navigation is used to locate the sorting area and complete operations such as finding, identifying, grabbing, and putting back items. The actual test results show that the robot has an average item recognition rate of 99% and a classification rate of 97.25%, which verifies the feasibility of the robot and provides a reference for research on intelligent collaboration, item classification, fast logistics, and machine vision.

Keywords

• ssd target detection
• items sorting
• color navigation
• mechanical control