Design of FPGA-based LAMOST fiber position edge detection module

In the closed-loop fiber positioning control mode of Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST), there are stringent requirements for time efficiency. However, due to the large size and high resolution of the images required for fiber positioning, the data volume for a single image is extremely large, reaching up to 90 MegaBytes per image. The current time efficiency of the closed-loop control system has yet to meet the requirements. This paper attempts to implement the fiber positioning algorithm, the light centroid algorithm, for calculating fiber positions in an FPGA-based edge fiber positioner detection module. Optimizing this module can significantly save image transmission time. The article mainly utilized Vivado HLS to port the algorithm to an FPGA and achieved improvements to its implementation. To enhance the algorithm's computational efficiency, the labels were assigned to each light spot to calculate the centroid coordinate, which enabled more efficient operation on the FPGA. The results indicated that the new hardware architecture effectively have reduced transmission time, making the improved centroid algorithm more suitable for operation on an FPGA.