According to the actual requirements of real-time target tracking processing and display of short wave infrared camera with CameraLink interface, an embedded image processing system based on FPGA is designed. The LVDS and GTX ports of FPGA are used to realize CameraLink interface and 3G-SDI interface, it not only improves the integration of the system, but also saves the cost. The adaptive dual platform histogram equalization algorithm is adopted and implemented, which effectively improves the display effect and subsequent processing quality of short wave infrared image. The paper briefly analyzes the algorithm principle of adaptive dual platform histogram, puts forward the adaptive calculation method of dual platform value, studies the implementation method of the algorithm based on Xilinx's FPGA, and gives the solutions of key modules such as adaptive dual platform histogram equalization, image enlarge and 3G-SDI display in FPGA. The experimental results show that the designed embedded image processing system gives consideration to real-time and effectiveness, and realizes the functions of image amplification and high-definition display; The algorithm can calculate and adjust the upper and lower platform values of the histogram in real time according to the scene within 400 mu s, effectively enhance the contrast of the image and suppress the background noise. The system has the advantages of strong algorithm adaptability, simple implementation, high integration and low power consumption. It can also be effectively applied to the image acquisition, processing and display of medium wave and long wave infrared cameras.