Texture image identifying method

Abstract

The invention discloses a texture image identifying method. The method comprises the following steps of segmenting texture images; deblurring each texture image block; performing binary processing and morphological processing; calculating the center of gravity in an extracted texture area; using a projection counting module to count projection histograms respectively along a polar radius and a polar angle in polar coordinate images; performing Fourier transformation on the projection histograms; selecting amplitude-frequency part information to acquire characteristic vectors with rotation invariance and making the characteristic vectors serve as characteristic vectors of detected images; performing the above-mentioned steps on texture images to be identified and standard texture images to acquire characteristic vectors; making the characteristic vectors merged and calculating the similarity between the characteristic vectors; and judging whether the texture images to be identified are standard texture images according to a similarity threshold value. The method which makes the texture images first segmented and merged makes full use of the advantageous characteristic among texture image single characteristics. The texture image identifying accuracy is thus greatly improved.

Description

Technical field [0001] The invention relates to the technical field of image processing, in particular to a texture image recognition method. Background technique [0002] With the vigorous development of Internet shopping, there is an increasing demand for group recognition of pictures, images and videos. At present, the detection and recognition of texture targets is an important part of target detection. When the camera and the target are stationary, there is no interference, rotation, or blurring, so the task of detecting the target is relatively easy. However, when there is relative motion between the camera and the detection target and the motion cannot be estimated, there will be various interferences such as target disappearance, target rotation, affine change, and image blur in the collected target image. For example, in the traffic field, the detection and recognition of road signs, license plate numbers, signs and patterns are currently commonly used methods such as ...

AI Technical Summary

Problems solved by technology

However, when there is relative motion between the camera and the detection target and the motion cannot be estimated, there will be various interferences such as target disappearance, target rotation, affine change, and image blur in the collected target image.

For example, in the field of traffic to detect and recognize road signs, license plate numbers, signs and patterns, corner detection, SIFT feature matching, ORB descriptors and other methods are commonly used at present, but these methods have defects in highly dynamic symbol recognition. It makes the target of detection and recognition blurred, and it is easy to cause misrecognition and missed recognition.

Images