Noise generation model of low-light remote sensing camera based on physical mechanism

With the continuous development of the field of remote sensing, the field of space-based remote sensing is developing in the direction of all-sky time and intelligence. Since low-light remote sensing is used to detect ground objects under low illumination conditions such as night and dawn and dusk, it leads to the characteristics of low contrast, low brightness and low signal-to-noise ratio of remote sensing images. The low signal-to-noise ratio leads to a large number of complex physical noises which will drown the image features and seriously affect the recognition and interpretation of ground objects. In this paper, a real physical model construction method based on physical mechanism of low-light image is proposed, By solving the noise parameters of space-based low-light remote sensing camera, the specific noise distribution model is constructed, which is used to synthesize the training data of the training denoising network, Thus, it gets rid of the difficult and laborious calibration problem caused by the lack of space-based data. In addition, in order to simulate the real imaging scene of space-based remote sensing camera, a set of low-light remote sensing noise image data set based on real physical model is constructed for the first time, which effectively promotes the development of subsequent image denoising research.