A method of positioning and recognition of electronic scale characters based on deep learning

Abstract

Digital characters are widely used in fields such as industry, equipment display and detection. According to the application background of reading recognition of electronic scale in chemical experiment examination of middle school, in order to improve the application of artificial intelligence in the education field, meet the requirements of intelligent examinations, and save manpower, a character positioning and recognition method based on the combination of deep learning and traditional image processing is proposed, which can automatically and quickly identify the reading of digital tube display screen of electronic scale, and according to the test requirements and test sites can automatically determine whether the students’ weighing operation is correct. Firstly, the yolov3 target detection algorithm and deeplabv3plus semantic segmentation algorithm are used to locate the character area of the display screen, and then a single digital tube character is obtained by projection segmentation. Finally, the recognition algorithm based on support vector machine is used to recognize characters. The experimental results show that the method has a high accuracy of positioning and recognition, and can adapt to different lighting, digital tube size and character tilt angle, has certain anti-interference ability, meets the real-time requirements, and has high application value.