Currently, embedded cameras are gradually replacing traditional cameras and entering the public eye, finding applications in many fields such as public security criminal investigation, biomedical science, and cultural heritage protection. However, with the increasing demand for higher quality visual imaging, image sensors are evolving towards high resolution, high sensitivity, and high pixel bit depth. Ordinary embedded cameras can no longer meet the processing requirements for large data volumes and low latency. To address this, this paper researches and designs an FPGA-based MIPI interface embedded platform camera, conducting in-depth research on image acquisition, image processing, and image transmission to ultimately develop the complete machine. For the front-end image acquisition of the system, it supports multiple image sensors from Gpixel, such as GSENSE400BSI, GSENSE2020S, GMAX0806, etc., driving the image sensor to output LVDS data signals through a pre-configurable mode. Through careful research, a functional board separation hardware design was proposed. This allows compatibility with different image acquisition circuit boards while using only one image processing circuit board, ensuring system integrity and effectively improving system R&D efficiency. The intermediate image processing part of the system uses a Spartan-6 series FPGA chip. Real-time data parsing modules and real-time data stitching modules adapted for Spartan-6 were developed. To achieve better image effects, three image sharpening algorithms are integrated within the Spartan-6: Sobel sharpening, Laplace sharpening