Design Of The Inspection System For The State Of The Blanking Opening Of The Transport Belt Based On Deep Learning

In the complex industrial environment,equipment monitoring is a very important link.The traditional monitoring method usually uses manual monitoring method for supervision,but this method will lead to missing detection due to fatigue of personnel.Therefore,using intelligent video monitoring to replace manual monitoring not only reduces the false alarm rate,but also greatly reduces the cost.In this regard,this thesis has designed a system to monitor the working state of the Blanking opening.In the design process of this system,there are difficulties such as the inability to directly monitor the working state of the Blanking opening,the small and easily stacked material targets in specific areas,and the complex environment.In view of the above difficulties and the overall design of the system,the main research contents of this thesis are as follows:(1)Aiming at the problem of how to detect the state of the Blanking opening,a new detection method is presented.Because it is impossible to directly detect the inside of the blanking opening,a method to indirectly determine the status of the blanking opening by detecting the material target in a specific area is obtained after research.The difficulties encountered in the detection method are analyzed one by one,so that the subsequent algorithm design can be better solved.(2)Aiming at the design of the algorithm model for the detection of the state of the Blanking opening,first,the Yolo v4 model detection counting and S-DCNet divide and conquer counting are used to count the material targets,and the conclusion is drawn that the divide and conquer counting method is better than the detection counting method;Then,to prevent the single model from missing judgment and misjudgment,an improved detection method combining the target counting model S-DCNet network with the target counting model Yolo-v4 algorithm with the calculated area of the ground target is designed to complete the detection of the state of the Blanking opening;Finally,this thesis proposes a blanking opening judgment mechanism suitable for this scenario to ensure that the computer can independently complete the status detection of the blanking opening.(3)For the design of the overall system,after the completion of the development of the detection model,the overall design of the system needs to conduct software and hardware building tests for the real needs of the site.The system realizes the functions of automatic image acquisition,monitoring alarm,human-computer interaction,etc.The target detection model,target counting model and state determination mechanism are applied to the system.When the system detects that the Blanking opening is in an abnormal or warning state,the system will send an alarm in time,and automatically generate an alarm log to assist the supervisor in judgment.At the same time,the system can also check and compare the alarm logs after the event.In this thesis,the part from the belt to the blacking port in the mineral aggregate transportation process of a real enterprise is taken as the detection object,and the detection system designed is used to test it.After many tests,the system can complete the detection of the state of the Blanking opening,and the accuracy of the test results meets the requirements of the on-site transportation process,and has good reliability.