Dynamic Image Processing for Visual Robot Cutting Steel Plate

When intelligent robot with visual guiding is cutting steel plates, the reflection interference from the captured images is possibly leading to the failure of tracking, and therefore the image processing and analysis methods is require of high robustness. An algorithm of tracking image feature points with active structured light is proposed based on comprehensive applications with advanced template-matching technology and correlation of multi-feature points. This algorithm overcomes the disadvantages of traditional extraction methods of feature points which were easily disturbed by intense noises. The experiment results indicated that the utilization of multi-feature points matching algorithm with dynamic template can improve the precision and stability of extraction feature points in presence of strong reflection interference and satisfy the request of real-time tracking.