With the intelligent identity recognition, face detection and recognition technology has been greatly developed, due to the blocking, face posture, face expression, image is affected by lighting conditions and low image resolution factors, face recognition rate is not high, affecting its cost performance, leading to slow market promotion.Opencv image acquisition module is selected as MCU, and the system is designed to record faces and recognize faces of people in the video shot by the camera respectively, which can realize the face recognition and identity mismatch alarm function of people under test.The design uses the camera module system for image acquisition, get different expressions and different state images, through the local binary mode histogram LBPH algorithm processing, further face facial feature extraction and comparison, so as to get the feature difference degree, face recognition. This system has been verified for many times and realized face recognition and discrimination functions, with high sensitivity, strong stability and other characteristics.