Adaptive structured light stripe center extraction method and system based on linear stretching

Abstract

The invention discloses an adaptive structured light stripe center extraction method and system based on linear stretching, and the method comprises the steps: carrying out threshold segmentation of a structured light stripe image f according to the average gray value gm and mean square deviation gs of the structured light stripe image f, and obtaining a segmented image f1; eliminating isolated pixel points in the segmented image f1 to obtain an image f2 after elimination; performing linear stretching on the rejected image f2 to obtain a stretched image fh; performing adaptive denoising on the stretched image fh by using a Gaussian filter to obtain a denoised image fg; and extracting a sub-pixel center point coordinate Pc (xc, yc) of the stripe center of the denoised image fg by adopting a square weighted gray gravity center method, and outputting the sub-pixel center point coordinate Pc (xc, yc). The invention has the advantages that the processing speed is higher, the adaptability is higher, the weight near the center of the stripe can be effectively increased, the weight is mainly concentrated on the pixel with the highest gray value, and the extraction precision of the center of the stripe is further improved.

Description

technical field [0001] The invention relates to the technical field of machine vision, in particular to a method and system for extracting centers of adaptive structured light stripes based on linear stretching. Background technique [0002] The structured light visual measurement system is widely used in road surface defect detection, medical diagnosis and industrial object defect detection with the advantages of active controllability, high precision and high real-time performance. One of the key technologies in the system, the accuracy of extraction directly affects the accuracy of the entire measurement system. [0003] In industrial applications, the light bands of line-structured light produced by lasers often show uneven brightness distribution and uneven width distribution due to environmental factors, which increases the difficulty of strip center extraction. [0004] At present, the methods widely used in extracting the center of structured light stripes include e...

AI Technical Summary

Problems solved by technology

[0007] According to the above formula, the disadvantages of the gray-scale center of gravity method are summarized into two aspects: one is that the extraction accuracy is low when there are unconnected highlight points in the same column of pixels; When the degree value difference is not obvious, the center extraction accuracy is reduced

[0008] In the improvement of the existing technology, on the one hand, the original gray value is directly added to the difference with the average gray value to construct a linear relationship to carry out the gray value stretching process, and then the Steger algorithm is used to extract the midline, reaching a certain Extraction accuracy, but the effect of linear stretch processing is not obvious and the parameters cannot be determined adaptively, and the algorithm speed is difficult to meet the real-time performance; The gray-scale center of gravity method calculates the center position of the light band, thereby obtaining higher accuracy and adaptability, but repeated threshold selection and Gaussian fitting lead to insufficient real-time performance, and the light band is prone to disconnection problems after binarization

Images