Abstract: In the vibration environment, the target video acquisition often be disturbed by the out factors, leading to image acquisition and poor stability, need to use image electronic image stabilization, traditional electronic image stabilization technology uses block matching method, by the shadow and obstacles occlusion interference, block matching electronic image stabilization effect is not good. An angle points gray compensation of image stabilization technique is proposed based on gauss gray compensation difference. For global motion estimation, the spatial neighborhood information to detecting motion of image in the magnitude of the amplitude characteristics, extraction of moving images, the image structure and texture information of analogy as a global movement of 3D RGB bit plane with the airport, the difference of gray level compensation, appropriate selection of the difference in different scales combined function to describe feature points of the image, pixel disparity target analysis and pixel image registration, achieve the corner point registration and electronic image stabilization. The simulation results show that, using the algorithm of image stabilization on instability of image, it can effectively improve the electronic image stabilization effect, the peak signal to noise ratio is improved, image stabilization performance is superior compared with the traditional method. The algorithm has important application value in the field of remote detection and target recognition.