Research And Development Of Embedded Vision Sensor For Two-dimensional Barcode Localization

With the rapid development of information technology and industrial automation,mobile robots are becoming more and more widely used.At present,mobile robots are widely used in many fields such as industry,power,manufacturing,service,medical and logistics.They can expand human work capacity,improve production efficiency,promote information management and production,and reduce or replace heavy labor of human beings.Positioning in mobile robot technology is one of the core technologies.The current common technique for solving the problem of mobile robot positioning is the two-dimensional barcode positioning technology.In this paper,an embedded two-dimensional barcode positioning sensor based on Xilinx's scalable processing platform Zynq-7000 was developed for the needs of indoor mobile handling robots.The function and performance indicators of the sensor system are determined according to the working scene of the indoor mobile robot and the electrical specifications of the sensor.The DM(Data Matrix)code is used as the sensing target,and the MT9V034 is selected as the image sensor.In this paper,the hardware electronic circuit,software,algorithm and structure of the sensor are designed.The hardware,structure,proofing,assembly,calibration and testing of the sensor are completed.The Linux system is transplanted to the Zynq-PS part,and the Zynq-PL FPGA development,software and algorithm development are also completed.At the same time,the two-dimensional barcode pattern of the sensor is designed and proofed.This design focuses on the analysis of filtering,binarization,edge detection,feature recognition,mapping transformation,positioning and decoding of DM code images.In order to increase the robustness of the algorithm in the feature recognition stage,a two-dimensional code localization algorithm is based on model constraints and data dimensionality reduction methods,then a simple and efficient corner detection algorithm is proposed.Because the image processing algorithm has natural parallelism,this design makes full use of the respective characteristics of Zynq chip PS and PL,and realizes the hardware acceleration of Sobel edge detection algorithm through reasonable allocation algorithm and software and hardware co-design,which ensures the development efficiency also reduces the power consumption and cost of the sensor.