A real-time blind pixel correction method for large array space remote sensing cameras in orbit

Large array remote sensing camera detectors have a large number of blind pixels, which is difficult to correct in real-time in orbit. It is difficult to meet the requirements of pixel-level image processing algorithms such as image feature extraction, registration and fusion. In this paper, based on the integrating sphere radiation calibration system, a method suitable for real-time blind pixel correction of large array space remote sensing cameras in orbit is proposed to solve the problems such as a large number of blind pixels, limited in orbit hardware resources, and high difficulty in real-time correction of high frame rate blind pixels. This method firstly performs blind pixel calibration on the ground. In orbit, a "3*3 sliding window" is used to calculate the address information of each pixel in real-time, as well as the distribution of blind pixels around the pixel and the replacement value of detector response. Determine the output original value or substitute value based on the segmentation of blind pixels around the pixel, and complete blind pixel correction. This method can take into account the situation where the blind pixels are located at the edge of the detector. This method has high reliability, less consumption of hardware resources in orbit. This method has no need for image download, and can match high frame rate image output. It has important reference significance for improving the temporal and spatial resolution of the large array space remote sensing cameras and improving the signal-to-noise ratio of the output images.