Research On Recognition Method Of Hazardous Chemicals Based On QR Code

With the rapid development of today's society and the rapid improvement of scientific and technological levels,hazardous chemicals have begun to occupy an important position in our daily lives.The high-precision identification of hazardous chemicals has become a hot point for experts and scholars.However,hazardous chemicals in complex and uncertain environments often cannot be quickly and accurately identified,which raises huge problems for the use and management of hazardous chemicals.With the help of rapid development of computer science,automatic recognition technology is receiving more and more attention from scholars.Fast and accurate automatic recognition technology is playing an important role in plenty of areas.Among them,QR Code must be specially mentioned.QR Code is a kind of two-dimensional code.Because of its low cost,large storage of information,strong error correction capability and easy sharing,QR Code is being widely used in numerous applications such as social software,positioning and navigation,cargo information tracking,warehouse management and objects recognition.This article will establish ' a hazardous chemicals identification algorithm based on QR Code and will mainly carry out the following research work:1.Due to the previously proposed atmospheric scattering model and convolutional neural network,an image enhancement algorithm based on Contrast-Limited Adaptive Histogram Equalization(CLAHE),the image dehazing algorithm of the modified Gated Context Aggregation Network(GCANet)and the improved adaptive thresholding algorithm are combined to recognize QR Codes in complex hazy environments.In order to evaluate the proposed algorithm,QR Code images under hazy conditions were collected and an image dataset was established.Then QR Code recognition accuracy using several algorithms was compared.The results show that the proposed algorithm can effectively improve the recognition accuracy of QR codes in environments with higher smoke density,and it is robust in complex hazy environments.2.Using the previously proposed Generative Adversarial Nets(GAN),a QR Code recognition method combining a high quality image deblurring algorithm based on GAN and an improved local adaptive thresholding algorithm is established.This method can be used in complex and uncertain motion blurry environment.It is robust in the environment with severe motion blur.In order to evaluate the proposed algorithm,QR Code images affected by motion blur provided by both public and our own database are selected as test set.The proposed algorithm can achieve higher recognition accuracy than other methods.It can maintain high recognition accuracy even when the image is severely affected by motion blur.3.For the edge detection and geometric correction problems of QR Code images.First,the mathematical morphology expansion method is applied to obtain the edge of the target area.Then the boundary of the target area is extracted by Canny operator.After that the boundary of the target area is transformed by Hough transform.Finally,QR Code images are corrected by applying bilinear interpolation operation.