Research On Binary Image Recognition In 2d Image Recognition Under Complex Illumination

With the rapid development and application of computer technology,barcode identification technology is increasingly used in manufacturing,communications,transportation,biomedical,fast payment,logistics management and other industries.Among the currently used barcode types,one-dimensional barcodes have the disadvantages of small information capacity,poor error correction capability and low space utilization.In order to meet the application requirements of barcode technology,The 2D barcodes have the characteristics of large information capacity,wide encoding range,strong error correction capability,and easy production,and have begun to be widely used in all walks of life.The two-dimensional codes of different code types have different characteristics,among which the QR code is widely used in the bar code market by virtue of its advantages of quick reading(360°),small occupied space,large data density and unique coding characteristics.When using an image sensor to collect barcode images,due to the influence of background interference,strong light,low light,and uneven illumination,the collected QR code image is prone to noise or the barcode area is not clear,such as overall highlighting and sudden brightness changes.With the gradual change and low contrast,it is difficult to select an appropriate threshold when the image is binarization resulting in problems such as low success rate of the QR code recognition and long time consuming.Therefore,we need to carry out detailed research on QR code image binarization to determine a binarization method that can accurately divide the bar code image.In this paper,the classification of QR code images and image binarization in the process of QR code recognition are studied.The main contents are as follows:Firstly research the bipolarization of complex light QR code images,analyze the characteristics of the collected QR code image samples,classify these images,and filter the image to reduce noise,then complex lighting for QR codes.In the image,an improved local threshold adaptive binarization method is proposed,and the complex light image of QR code is analyzed experimentally.The result shows that the success rate of the image after the method is 93.33%.Identifying open source code bases based on Google barcode,completing the development of QR code recognition system in Linux environment,and successfully transplanted on the ARM9 hardware platform,completed a 2D barcode reading device,and the experimental results show that the reading device The QR code can be accurately read under complex lighting conditions,and the success rate is above 80%.In this paper,we study the image binarization and decoding technology through the complex light image of the QR code collected by the 300,000 pixel image sensor.Based on the embedded Hi3518EV200 hardware platform,a 2D barcode scanning device is implemented.Based on the hardware and software platform,the code is optimized and designed,and the recognition success rate of the QR code identification algorithm is improved to more than 80%.And for strong light,low light,uneven lighting and other complex lighting conditions,the scanning device can complete the QR code image acquisition,image processing,decoding,decoding and information display and other functions.