In order to realize high-speed motion detection and camera ego-motion estimation in image processing，an insect-inspired Reichardt motion detector (Elementary Motion Detector (EMD)) and receptive fields based on insects vision system are applied.The principal characteristics and drawbacks of the Reichardt model are analyzed.According to one of its main drawbacks，a modified model is selected which performs better in motion detection than the original method.Moreover，six templates of receptive field based on flys vision system are designed for simple ego-motion estimation，such as self-translation and self-rotation of the camera.Finally，the related algorithms are implemented on a FPGA (Field Programmable Gate Array) platform.The results of several typical experiments demonstrate，that the EMDs can detect local optical flow quickly and the receptive field templates enable simple motion estimation under specific backgrounds.The developed FPGA system is sufficient to deal with a video frame rate of 350 fps at 256×256 pixels resolution，the resulting time delay of the Reichardt model implementation is only 0.25 μs.This hardware can be applied to real-time computer vision systems，such as for obstacle detection，motion estimation，UAV/MAV’s stability control and so on.