Aiming at the problems of low efficiency of manual verification methods and high development cost of reading recognition system when the measurement verification laboratory verifies meters, a Raspberry Pi-based automatic measurement verification laboratory meter reading recognition system was designed. The system uses the Raspberry Pi 4B to control the camera to collect the picture of the meter reading, and then uses the SSDA algorithm to locate the image reading area, and perform character segmentation, and then use the BP neural network and feature matching algorithm to recognize the segmented characters, and finally use the SSH method to recognize the results Send to PC for storage. The experimental results show that the running time of the system is 500-600ms, and the accuracy of reading recognition is above 92%, which proves that the system can be used for automatic recognition of meter readings in metrological verification laboratories.