Industrial Video Surveillance Image Denoising based on Non-subsampled Contourlet Transform

With the continuous expansion of industrial video scale, visualization technology has become an indispensable technical means. However, during the collection and transmission of images, they may be affected by factors such as the device itself and environmental factors, which can lead to noise pollution and a decrease in image quality, which is not conducive to image post-processing. Therefore, industrial video surveillance image denoising has become an important research topic, and using various methods to improve image quality has become an important research direction. In order to address the ringing effect caused by hard threshold denoising, this paper proposes an adaptive smooth non-subsampled contourlet transform (NSCT) denoising algorithm. The proposed method utilizes the correlation of each sub-band to improve the denoising function, making it adaptive. The simulation results show that the algorithm proposed in this paper improves subjective perception such as contrast and detail clarity of remote sensing images, as well as objective evaluation indicators such as peak signal-to-noise ratio, information entropy, and correlation coefficient