Target Detection Method Based on Joint Matching of Edge and Region Features

In the field of robot autonomous recognition and grabbing of objects, this paper proposes a method based on joint matching of edge and region features. First, this method performs median filtering and noise reduction preprocessing on the collected images, extract object edge information using Canny edge detection algorithm, and then segment the image area using edge information. Secondly, use the target object template to perform the first contour moment matching with the segmented edge area set image to obtain the candidate area set image of the target object. Finally, the target object is identified by matching the feature points in the template image and candidate region set image. The differences between traditional methods and this method under low light conditions are compared, the experiment have shown that this method can detect and recognize target objects under darker conditions, with a success rate of over 94%, greatly improving the accuracy of object detection in the scene mentioned above.