A Target Recognition Method Based on Image Contour Features

Abstract

The invention discloses a target recognition method based on image contour features, which preprocesses a template image and an image of an object to be measured to generate a binary image; establishes a feature library of the object template contour: extracts the complete contour of the binary image of the object template, and generates a binary image in the contour Take a certain number of feature points at equal intervals, and use the context features of the feature points to describe the object outline; perform target recognition on the image of the object to be detected: extract the contour edge of the binary image to be detected; and select certain feature points; The shape direction of the object is converted into the orientation of the template; the context feature of the selected point is used to describe the contour of the object to be tested after the orientation is transformed; the similarity between the object to be tested and the template object is measured by matching cost. Compared with the prior art, the invention solves the problem of rotation invariance in the contour matching process, makes the rotation in the contour matching process invariant, and is effectively applied to target recognition in images.

Description

technical field [0001] The present invention relates to the technical field of machine vision and pattern recognition, and more specifically, to a target recognition method based on image contour features. Background technique [0002] The current target recognition methods can be divided into three categories: methods based on appearance features, methods based on shape features, and methods based on the combination of appearance and shape features. Object recognition based on shape features is a hotspot in current research, and significant progress has been made in recent years. From the traditional methods based on Fourier transform or invariant moments, it has developed into the current shape matching methods that are mostly based on the two shape descriptors of contour and skeleton. Compared with point sets, contours have richer information, and contours are not easily affected by changes in illumination, object color, and texture. The most important thing is that they...

AI Technical Summary

Problems solved by technology

[0005] However, none of the above methods can find a simple and effective description of the contour to solve the problem of rotation invariance in the contour matching process

Images