Stripe Noise Removal Method of Infrared Image on Satellite Based on Similarity Line Window Mean Compensation

Abstract

The invention discloses a method for removing streak noise in an infrared image on a satellite based on the mean value compensation of a similar line window. The method first selects a one-dimensional sliding window of size N according to the image size as a sliding line window. The sliding line window scans from the top end to the bottom end of each column along the direction of stripe noise, and selects the line window with the smallest variance, which is The similarity line window of each column; then, calculate the average gray difference between the similar line window and its adjacent area, and obtain the stripe noise difference value of the adjacent column; finally, set the stripe noise value of the first column to zero, according to the adjacent column Iteratively calculates the noise value of each column, and subtracts these noise values ​​from the original gray value to obtain the ideal gray value of the real object. Compared with the prior art, the method has low computational complexity, good robustness and high fidelity to texture details.

Description

technical field [0001] The invention relates to a method for removing streak noise of an infrared image on a satellite based on similar line window mean value compensation, and belongs to the technical field of digital image processing. Background technique [0002] During the formation of infrared images, noise exists in every link of image acquisition, such as periodic offset of sensors, electromagnetic interference of load components, etc. The generation of fringe noise is caused by the difference in response to repeated scanning of the scanned detection element during the imaging process of the satellite’s repeated scanning of ground features by sensors and photoelectric devices. Different photosensitive elements will produce different output signals for the same infrared radiation, and the sensor scanning A special noise with periodicity, directionality, and stripe distribution caused by disturbances caused by various factors such as mechanical movement. [0003] Infra...

AI Technical Summary

Problems solved by technology

However, the adaptability and universality of these methods are poor, and often require a single type of object in the image, and the algorithm calculation is complex

For such situations, many scholars have proposed improved algorithms for streak noise. Although these methods achieve the purpose of maintaining image quality and effectively removing streak noise to varying degrees, the effect is not ideal.

Images