Research On High-speed Readout Method Of SCMOS Camera With Large Dynamic Range

The High Energy Cosmic-Radiation Detection Facility(HERD)project is a planned large-scale scientific facility of the Chinese Space Station.The purpose is to carry out dark matter particle detection tasks and solve the dark matter low-light ultrafast detection system technical challenge.Because of its large dynamic range and high sensitivity,the sCMOS camera has attracted the attention of scholars in the field of low-light detection.In order to realize the ultrafast low-light detection of dark matter with a large dynamic range,the development of a large dynamic range enhanced sCMOS(IsCMOS)camera system needs to be completed.This article focuses on the needs of the space high-energy cosmic radiation detection facility project,and develops a large dynamic range and high-speed sCMOS signal measurement technology output by the calorimeter.The main research work is as follows:1.Using MATLAB to program the process of reproducing the output signal of the calorimeter.The study shows that the original signal is modulated by the image intensifier and light cone in the detection system.The signal form will be converted from a pulse signal to a Gaussian signal.Through experiments,it is found that when in the far field,the output signal is Gaussian function.Because the output signal form is basically the same,this paper simplifies the complex process of calorimeter output to obtain Gaussian function of the signal through multimode fiber under the condition of far field.We can get the core algorithm by analyzes and processes the data under the premise of Gaussian function;2.The sCMOS detector can adjust two amplifiers and output two images with different gains at the same time,which has high sensitivity.In this paper,according to the unique output characteristics of the sCMOS detector,an algorithm based on feature point fusion is proposed to achieve The purpose of expanding the dynamic range of the detector is to ensure that the detector can detect the signal under weak light conditions;3.The data format of each image output in this project is that the 16-bit image with a frame rate of 500 and a size of 20 * 20 is output under a 7500-channel optical fiber.Its bandwidth is about 20 Gb / s,but when transmitting data,the bandwidth is only 1Gb / s,so in the face of such huge data,data compression is very necessary.This paper analyzes the characteristics of the Gaussian function,and proposes two algorithms that can quickly read the output signal of the calorimeter.They are the transmission of selected feature points based on the Gaussian Newton fitting method and the feature function method based on the multidimensional Gaussian function.To compare and evaluate the characteristics and effects of these two methods,it is found that the Gaussian-based feature function algorithm has a good compression rate and can achieve a good fitting and restoration effect when the signal is in the unsaturated state,but the feature is saturated The point extraction method has better fitting and restoration effect than the feature function method,so this paper designs a hardware implementation process that fuses the two algorithms together.Under the condition of ensuring a certain compression rate,the error is minimized as much as possible to meet Requirements for fast reading algorithms.