A Vision-Based Measurement and Classification System for Robot Arm Under Controlled Lighting Condition

This paper presents development of a measurement and classification system for robot arm using machine vision under controlled lighting environment. The proposed system uses a single camera as a sensor for measuring and classifying objects which are bolts and nuts. Using image processing and analysis, characteristics of objects was extracted and area of blob in binary image also was calculated for classification process. For coordinate calibration process, the quadratic transformation and regression analysis were used to determine relationship between image coordinate and the world coordinate. Experiment results showed that the proposed system can measure and classify the components exactly of 100% from all samples tested and measurement errors are suitable with the system which applied to a robot arm.