Method, apparatus and electronic device for detecting dirty areas in an image

Abstract

Disclosed are a method, apparatus and electronic device for detecting dirty areas in an image. The method includes: performing a first scale Gaussian filter on an image to be detected to obtain a background image corresponding to the image to be detected; Making a difference between the image to be detected and the background image to obtain a difference image; Performing a second scale Gaussian filter on the difference image to obtain a resultant image, the scale of the second scale Gaussian filter being smaller than the scale of the first scale Gaussian filter; And determining a dirty area in the image based on the resultant image. In this way, the sensitivity, ease of use, processing speed and environmental adaptability of dirt detection can be improved.

Description

technical field [0001] The present application relates to the field of image processing, and more specifically, to a method, device and electronic equipment for detecting dirty areas in an image. Background technique [0002] At present, in many fields such as computer vision, automatic driving and security systems, and industrial visual inspection systems, it is necessary to use camera modules to collect images. [0003] During the production and use of the camera module, the imaging surface of the module will be dirty due to various factors (touch, scratches, abrasions, etc., stains, pixel damage, etc.), which will affect the product quality and user experience of the module. . [0004] Since the shape of the dirt on the camera module is arbitrary, the gray level difference between the dirt and the good image is small, which makes the detection of the camera module very difficult. At present, most module manufacturers use manual inspection for camera module dirt detectio...

AI Technical Summary

Problems solved by technology

[0005] However, there are obvious defects in the manual detection method. Due to the influence of the inspectors' own factors and the differences in the test environment, the false positive rate of the detection is high, and the consistency of the detection results is poor. Some inconspicuous dirt may even be missed.

Images