Low-light Image Enhancement Model with Low Rank Approximation

When taking pictures in low light or backlight, the obtained image often appears too dark or uneven light distribution, resulting in poor visual quality of the image. The dark light enhancement model based on the Retinex model can achieve effective light enhancement. However, such dark light The enhancement model also has some problems, that is, although the visibility of the dark area in the image to be processed has been effectively improved, the hidden noise is also amplified and highlighted, which still affects the visual quality of the enhancement results. To solve this problem , a dark-light image enhancement model based on low-rank matrix estimation is constructed. First, a Retinex model containing noise terms is constructed and optimized alternately, and the dark-light image is decomposed into the illumination layer I and the reflection layer R. In this process , the noise suppression of the R layer is realized by using low-rank matrix estimation. Secondly, considering the problem that the image detai