Research On The Extraction And Recognition Algorithm Of Industrial Twodimensional Code Based On Embedded System

Two dimensional code has been widely used in life and production, which has the characteristics of high-capacity, strong anti-wear performance and so on. Especially in the field of industry, the use of two-dimensional code recognition has become an industry standard of automation industry which is used to tracking the workpiece of production, assembly management, life cycle maintenance and so on. Two dimensional code is different from common printing, the application environment of industrial two-dimensional code is extremely harsh, the twodimensional code recognition is usually accompanied by noise, exposure, wear and other issues. In particular, the identification of direct component identification(DPM), the conventional two-dimensional code recognition method can not meet the needs, and the domestic industry is not mature competitive two-dimensional code identification products. So it is important and urgent to study the technology of twodimensional code identification in industry.In this paper we make a deep research on industrial two-dimensional code extraction and identification methods base on DSP platform for above demands, design and improve the extraction and recognition algorithm of Quick Response and Data Matrix in industrial environment, and verify the reliability of algorithm by hardware to achieve the expect results. The main research contents and results of this paper include the following aspects:1. Making a deep research on types and technology features about two-dimensional code. Analyze and improve the extraction algorithm on Quick Response and Data Matrix codes. Implement and improve the traditional two-dimensional code extraction process as well. Otherwise, we make comparison and verification on variety of algorithms in key steps. The experimental results show that the algorithm has better robustness and higher efficiency;2. The actual DPM will defaced or destroyed in the circulation process so we improve and optimize the algorithm to accommodate these industrial twodimensional code recognition demands. We also improve the algorithm by different application scenarios test in order to achieve a more general and high performance;3. The algorithm is deployed to DSP platform to fit the real industrial demands. According to problems encountered in the process of transplantation such like compatibility and performance optimization, we make a detailed introduction on DSP platforms and official library functions in this article. In addition, a PC host computer software is developed and introduced as well. The recognition rate of this algorithm for two-dimensional code made on the surface of metal can reach 95%, the recognition rate of printed two-dimensional code can reach 98%, the average recognition time of QR code is 60 ms, which the average time wasted by DM code is 80 ms.