Research of digital meter identifier based on DSP and neural network

Based on image recognition and wireless short-range transmission, we designed and implemented a kind of equipment to identify the value of several digital domestic instruments such as kilowatt-hour meter and gas meter. CMOS image sensor is used to form focused image of the meter reading. Then, the digital image is processed by a high-performance DSP processor with advanced neural network algorithm. The identification result is specific numbers and is sent to the concentrator through the short-range wireless ZigBee communications. The equipment can effectively release the workers from entering residential buildings and avoid many troubles.