Warehouse box body identification and positioning method based on contour features

Abstract

The invention provides a storage box recognition and positioning method based on contour features. The method comprises the following steps: (1) collecting a color image and a depth image of box goods; (2) performing foreground segmentation on the color image and the depth image, mapping depth blocks to the color image, marking color region blocks falling into the same depth block, and combining the color region blocks to serve as a final box cargo foreground image; (3) performing linear edge detection on the foreground image of the box cargo, and connecting edges by taking a rectangle as an outline target to obtain a candidate target closed outline; (4) selecting the target meeting the predetermined box body rectangular characteristics from the candidate target closed contours to obtain abox body identification result. The method performs identification and positioning of the box goods based on the depth image and the color image, improves the calculation efficiency, has the advantages of low cost and high real-time performance, and is suitable for identification and positioning of the box goods in complex environments such as factory warehouses.

Description

technical field [0001] The invention relates to the field of identification and positioning of machine vision, in particular to a method for identifying and positioning goods in a box based on a depth camera. Background technique [0002] At present, most of the existing multi-camera-based object recognition and positioning equipment use stereo matching images collected by different sensors to obtain the position of each pixel in space, which will lead to high cost, slow operation speed, and too complicated system issues such as sex. [0003] Most of the object segmentation is realized by the method of convex hull extraction based on color images. This processing method needs to consider the surface color and texture of the object. When the background color is similar to the object, it is easy to cause misjudgment. Obviously, it is difficult to capture its feature points with a general color camera. Contents of the invention [0004] Aiming at the above defects or improv...

AI Technical Summary

Problems solved by technology

[0002] At present, most of the existing multi-camera-based object recognition and positioning equipment use stereo matching images collected by different sensors to obtain the position of each pixel in space, which will lead to high cost, slow operation speed, and too complicated system sexual issues

[0003] Most of the object segment

Images