THE RESEARCH AND DEVELOPMENT OF ROBOTIZED VISUAL TRACKING SYSTEM FOR DETECTION OF SPIRAL SEAM IN LARGE DIAMETER STEEL PIPE

This paper introduces a visual-feedback-based moving seam tracking system which is appllied to the automatic line of ultrasonic detection of spiral-seam-pipe. The system processes real-time pattern recognition and system control by software running on a single CPU. In this paper.we propose a matching recognition stratrgy to establish feature model on-line and to process pattern recognition with this model.The system is adaptive to the seam feature changes that are caused by several reasons such as changes of steel pipes,work environments and conditions. It can recognize seams with certainty and rapidity.The system displays good tracking performance on real detection line.