A Fabric Defect Detection Method Based on Light Field Camera Depth Information Extraction

Abstract

The invention relates to a fabric defect detection method based on light field camera depth information extraction, the method can be used in the field of fabric detection, and is a method for detecting fabric defects from a 3-dimensional space. This method utilizes a light field camera to generate a multi-view image sequence of the target fabric. By substituting the slope for the estimated depth value of the disparity, the depth map is obtained by using the method of sub-pixel offset. In order to avoid the interference of the vignetting effect of the microlens of the light field camera, the weakly affected area is extracted, and the noise reduction process is performed on this part. The noise reduction of the present invention uses an image self-adaptive window filter noise reduction method. Effectively avoid the error caused by too large and too small median filter window. Finally, binarization is performed to obtain a segmentation map. By using the method of the invention to process the fabric, the defective part of the fabric can be effectively detected.

Description

technical field [0001] The invention relates to a method for detecting fabric defects, especially the processing, extraction and detection of three-dimensional depth information of fabrics Background technique [0002] The earliest depth processing is to use the camera array to obtain the parallax. The light field camera replaces the complex camera array with the main lens and the microlens array. One camera can realize the function of a camera array to obtain the depth. Most of the existing light field camera depth processing technologies deal with some relatively large objects, or the shooting target is far away from the imaging surface, such as the depth processing of light field cameras used for statues and Lego cars, because the light field camera microlens array Because of the vignetting effect, there will be a vignetting effect, which will have a great impact on close-range high-precision and high-texture complexity targets, such as fabric textures. The current light...

AI Technical Summary

Problems solved by technology

Most of the existing light field camera depth processing technologies deal with some relatively large objects, or the shooting target is far away from the imaging surface, such as the depth processing of light field cameras used for statues and Lego cars, because the light field camera microlens array Because of the vignetting effect, it will have a great impact on close-range high-precision and high-texture complexity targets, such as fabric textures, etc.

The current light field camera depth map construction technology has been very developed, and the light field camera can be used to realize 3D reconstruction very clearly, but the vignetting effect of the light field camera has not been well solved yet.

Existing solutions are all estimated based on the particularity of the object, and the estimated results cannot be used in the detection field

Images