A Method of Image Singular Point Correction

Abstract

The invention discloses an image singular point correction method, belonging to the image signal processing field. The method includes finding out singular points in an input bayer image through a singular point detection module; judging the type of the singular point according to a singular point type judgment module; the singular point correction module is used according to different types; thesingular points are corrected by adopting different methods. According to the method, the singular points in the image are corrected, the impression of the image can be improved, the singular points are detected by adopting a difference value between the singular points and surrounding sample points, a special environment and a special procedure are not needed, the implementation complexity is greatly reduced, the singular points are corrected, and the definition of the image cannot be influenced.

Description

technical field [0001] The invention relates to a method for correcting singular points, and more specifically, relates to a method for correcting singular points of an image, which belongs to the field of image signal processing. Background technique [0002] The image singularity point refers to some sample points in the image that are too different from the surrounding pixels, which is visually harsh. According to the formation reason of the singular point, there are generally two ways to correct the singular point in the prior art. Specifically: the first one is that due to the manufacturing process and material differences, some pixel positions of the image sensor are photosensitive, that is, no matter what the external light is, the position always displays the same brightness. The characteristic of this singular point is in the image The position is fixed. According to its value, it can be divided into two types: "bright spot" and "dark spot". It can be stored in adv...

AI Technical Summary

Problems solved by technology

Specifically: the first one is that due to the manufacturing process and material differences, some pixel positions of the image sensor are photosensitive, that is, no matter what the external light is, the position always displays the same brightness. The characteristic of this singular point is in the image The position is fixed. According to its value, it can be divided into two types: "bright spot" and "dark spot". It can be stored in advance and then corrected in a special way, but storing the singular point position requires a lot of memory space , especially for images with larger resolutions, the implementation cost is higher. At the same time, the position of each photosensitive component is different, so it needs to be detected in a special environment, and the process is more complicated; the second is that the singularity is caused by noise. The value and position of the singular point in the image are not fixed. In actual processing, it is usually corrected by filtering and using the weighted value of the surrounding samples, which will affect its clarity.

Images