Reserach On Online Inspection System Of Emulsion Explosive Packaging Defect Based On Machine Vision

The roll packaging is extremely important as the last step in the quality assurance of emulsion explosives.Due to the complex and varied environment,the emulsion explosives may have various defects in the packaging process.How to efficiently and accurately screen the defective drug rolls is a difficult problem.In order to solve the problem of packaging defects of emulsion explosives,this paper designed a machine vision-based medicine package inspection system.The main research content of the subject consists of five parts:(1)Firstly,the packaging process of the emulsion explosive packaging production line is introduced in detail.Then,the defects of explosives that may occur in the packaging process are summarized.The design idea of using the machine vision system to ensure the reliable operation of the emulsion explosive packaging production line is proposed.The image is grayed out by the optimal threshold method,then the image is processed by filtering wave,and the hough transform is used to extract the edge of the emulsion explosive,and the method of canny,laplace,etc.is compared.The superiority of the hough transform is highlighted.(2)The workflow of the machine vision system is introduced.The modular design method is adopted.According to the design requirements and the expected detection results,the function division of each module and the required effects are elaborated.The overall design of the machine vision system is proposed,and the software and hardware design of each component of the machine vision system is completed.In the hardware design process,the equipment was selected according to the environment of the explosive packaging production line and the system inspection expected to achieve the target accuracy;in the software design,the overall function and interface operation of the system were first introduced.Then the experimental verification of the effectiveness of the system shows that the system can efficiently and accurately detect the emulsion explosive packaging production line online.(3)The image of the drug roll is processed using an optimal threshold and smoothing method.Then the Sobel operator is used to extract the edge of the emulsion explosive,and compared with the edge detection methods such as canny and laplace.The comparison results highlight the superiority of the Sobel algorithm.Finally,the outline of the medicine roll is drawn,the morphological characteristics of the medicine roll are obtained,and the defective medicine roll is detected by the template comparison method.(4)A BP neural network classification method based on genetic algorithm optimization is proposed for the classification of the target of the drug roll.The shape feature of the target is extracted as the input of the network,and the foreground target category is used as the training result,and the model of the classifier is obtained through continuous training.The experimental results show that the classification effect is more significant than the BP neural network alone,which can effectively help the operator to obtain the status information of the production line.(5)The overall design of the machine vision system was proposed,and the software and hardware design of each component of the vision system was completed.In the hardware design process,the equipment was selected according to the environment of the explosive packaging production line and the system inspection expected to achieve the target accuracy;in the software design,the overall function and interface division of the system were first introduced.Then the experimental verification of the effectiveness of the system shows that the system can efficiently and accurately detect the emulsion explosive packaging production line online.