Research On Container Damage Risk Assessment And Detection Based On AHP And CNN

The container operation runs through the entire business process of the container terminal.The damage to the container surface,structure,and parts will cause the hidden danger of safety risks in the container operation process.With the continuous deepening of the automation of container terminals,j container terminals need more and more container operations to be completed correctly and efficiently.Therefore,it is more and more important to systematically study the impact of container damage on various operation links,and to discuss the pre-assessment and timely detection of container damage risks.Based on the existing research,this thesis studies the problem of risk assessment and damage detection of container damage.On the premise that container damage and its influencing factors are clear,the damage risk assessment model based on Analytic Hierarchy Process(AHP)is used to determine the types of container damage that need attention in different operations.On this basis,through the risk detection model based on a convolutional neural network(CNN),the identification of various damage types is realized.Therefore,according to the characteristics of different operation links,corresponding container damage risk management and control plans are formulated to achieve effective and reasonable container damage management.First of all,this thesis compares and summarizes the international container standards,container leasing standards,port container handover,and inspection regulations and other data documents,analyzes the subject and object,internal and external elements,and risk hierarchy involved in container damage,and constructs container damage risk assessment and detection.index system.Then,on this basis,a container damage risk assessment model was constructed based on AHP,and the coupling relationship between container damage risk factors and risk levels was further analyzed.Evaluate and propose risk plans based on the evaluation results.After that,a container damage risk detection model was constructed based on CNN,and multiple types of container damage data sets were constructed at the same time,and the identification and detection of various container damage types were realized through transfer learning and Mobile Net V2.Finally,the test is carried out in the operating environment of a container terminal in China to verify the feasibility and applicability of the container damage risk control plan and detection method.Combining theoretical research and practical verification,this thesis enriches the research on container damage risk management and contributes to the improvement of container damage risk assessment and detection methods.In addition,this thesis proposes a control and detection scheme for the risk of container damage,which provides a new idea for the risk management of container damage in container terminals.