Research On Portable Corrugated Board Counter Based On Machine Vision

The modern logistics and packaging industry has a large demand for corrugated board,In the face of the large amount of corrugated board produced by the production workshop every day,how to count the quantity of corrugated board quickly and efficiently is becoming an important problem faced by this industry.At present,the counting method of corrugated board usually has some problems,such as low efficiency,high error rate,high equipment cost and difficult maintenance.Because machine vision technology has the advantages of high real-time and non-contact measurement in industrial production,the method of counting corrugated board based on machine vision will gradually become the development trend in the future.In this thesis,a portable corrugated board counter based on machine vision technology is designed for the counting requirements of corrugated board stack in the industrial field.The details are as follows:(1)Design of imaging system for portable corrugated board counter.In this thesis,a portable corrugated board counter based on machine vision is designed according to the requirements of corrugated board counting accuracy in the industrial field and the portable requirements of the market for instruments,including the imaging system and the hardware structure of the counter.In addition,the appropriate hardware were selected according to factors such as the common height of corrugated board stack,the characteristics of corrugated thesis,the accuracy of detection and the demand for cost.(2)The counting method for the image of corrugated board stack based on hole feature model.Aiming at the problem that it is difficult to count the corrugated board in the image with blurred imaging and low contrast,this thesis proposes a hole feature model,which uses the image two-step information to enhance the holes in the image of corrugated board stack.Firstly,the corrugated board stack region is extracted according to the edge texture features of corrugated board,and then the scale maximum hole likelihood function is obtained in the scale space to enhance all holes in the image.Finally,the number of corrugated board is counted through the post-processing of all holes.In this thesis,experiments are carried out on many different types of images of corrugated board stack,the experiments verify the feasibility of counting corrugated board based on hole feature model.(3)The counting method for the image of corrugated board stack based on U-net model.When the environment where the corrugated board stack is located is complex,the counting method based on hole feature model may not be able to accurately extract the corrugated board area in the image.To solve the problem,this thesis uses U-net neural network to segment the holes of the corrugated board in the image.And a block method for the image of corrugated board stack is proposed,which can correctly block large-sized image of corrugated board before inputting into the model.Finally,experiments are carried out on images of many different types of corrugated board stack,and the feasibility of the method is verified by experimental tests.(4)Application software development and experimental verification.Firstly,the application software is developed and the whole counter is tested.Then the counting methods based on hole feature model and U-net model are used to count multiple groups of images of corrugated board stack.After the experiments,the counting accuracy,advantages and disadvantages of the two methods are compared and analyzed.Finally,how to combine the two methods in practical application is summarized.The portable corrugated board counter proposed in this thesis is mainly for the counting requirements of different lighting environments and different types of corrugated board stacks in the production workshop.After testing,it is shown that for corrugated board stacks with a height of less than 2m,the counting accuracy of the corrugated board counter proposed in this thesis can reach more than 97%,which meets the accuracy requirements of corrugated board counting in the industrial field.