Intelligent port container body damage checking method

Abstract

The invention provides an intelligent inspection method for harbor container body damage. The method comprises the steps of S1, judging whether a container transport vehicle exists in a designated position of the view of each camera or not, and if yes, turning to the step S2; if not, returning to the step S1; s2, controlling each camera to zoom the focal length to shoot the container transport vehicle in different directions so as to obtain container body images of different surfaces of the container body; s3, identifying each box body image to obtain box number information, and transmitting each box body image to an identification server for box body damage identification to obtain a corresponding damage identification result; and S4, the container number information and the damage identification result are linked with a tallying platform, and a worker at the gate is prompted to carry out manual intervention or release. The beneficial effects of the invention are that the system can achieve the automatic and intelligent container defect detection, and improves the defect detection efficiency and accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of damaged container identification, and more specifically, to an intelligent inspection method for damaged port container bodies. Background Art

[0002] With the development of the transportation and waterway industries, port operations (such as loading and unloading operations) have become increasingly frequent. To ensure the safety of cargo during transportation, the cargo is usually loaded into containers for transportation. Among them, a container is a large loading container with a certain strength, stiffness, and specifications and is specifically for turnover use.

[0003] Currently, when using containers to transport goods, the goods can be directly loaded from the warehouse at the loading port and unloaded after arriving at the warehouse at the unloading port. Even if the ship is changed midway, there is no need to take the goods out of the container for reloading. Generally, the detailed information of the container is recorded on the outer side of the container body, such as the container number, container size, container type, container function, loading and unloading status, special container mark, trailer number, and trailer container position number, etc. Among them, the container number is a code used to represent information such as the container owner, registration number, and check code. In addition, before transporting the goods, it is also necessary to create the ship chart data of the ship according to the detailed information of the loaded container.

[0004] Currently, the inspection of damaged containers at each terminal adopts an artificial method for verification, with low work efficiency and frequent missed inspections, resulting in terminal losses and affecting the damaged container replacement business. The inspection of damaged containers at the gate is a key link in the gate-in process. In the past, each incoming container had to pass through the pictures taken by the intelligent gate manually (one picture for each of the front, back, left, right, and top 5 faces), and the artificial judgment of the damaged condition of the container body was carried out. This link is also the most time-consuming link in the process of the intelligent gate entering the gate. At the same time, it may also cause damaged containers to be received due to different damaged judgment criteria of the gate personnel, resulting in economic losses. In the shore tallying link, there may also be different damaged judgment criteria for the gantry crane driver and the tally clerk, resulting in damaged containers being received and causing economic losses. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to realize automatic and intelligent container damage inspection, improve the inspection efficiency and accuracy. To overcome the defects of the above prior art (or related art), the present invention provides an intelligent inspection method for damaged port container bodies.

[0006] The present invention provides an intelligent inspection method for damaged port container bodies, which is applied to the tallying platform in the terminal tallying system. The terminal tallying system further includes an identification server. Multiple cameras in different orientations are installed at the gate of the terminal scene in advance. The intelligent inspection method for damaged port container bodies includes the following steps: Step S1, determining whether there is a container transport vehicle within the designated positions in the fields of view of the respective cameras: If so, proceed to step S2; If not, return to the said step S1; Step S2: Control each of the cameras to perform focal length zooming to capture images of different sides of the container body of the container transport vehicle from different orientations to obtain the body images of different sides of the container body; Step S3: Identify the container number information from each of the body images, and transmit each of the body images to the identification server for identification of container body damage to obtain the corresponding damage identification result; Step S4: Link the container number information and the damage identification result with the tallying platform to prompt the staff at the gate to perform manual intervention or release.

[0007] Compared with the prior art, a method for intelligent inspection of damaged container bodies in a port according to the present application has the following advantages: In the present application, the presence of the container transport vehicle is judged through step S1, the body images are captured through step S2, the damage of the container body is identified through step S3, and the container transport vehicle is processed through step S4 by linking with the tallying platform. Only by installing and using cameras can the tasks of detecting the container transport vehicle and obtaining the body images be completed. The task of inspecting the damaged container body is carried out through the identification server. Without affecting the container identification, the accuracy of damage identification is ensured, and the inspection efficiency can be improved without manual intervention. Without changing the current operation process and operation efficiency of the container terminal, the inspection of damaged parts is integrated. Finally, the tallying platform in the terminal tallying system and the gate are associated to realize automatic and intelligent inspection of damaged containers.

[0008] In a possible implementation manner, in the said step S1, the real-time images captured by each of the cameras are obtained, and it is judged whether the container transport vehicle exists in the specified positions within the corresponding fields of view in each of the real-time images: If so, proceed to the said step S2; If not, return to the said step S1.

[0009] In a possible implementation manner, in the said step S1, for each of the cameras in different orientations, the real-time image captured by each camera is separately divided into regions, the setting of the limited area of the container body is carried out, and the set limited area of the container body in each real-time image is used as the specified position within the field of view.

[0010] In a possible implementation manner, if the number of the cameras is 4 - 6, then in the said step S2, the body images of 5 sides of the container body are captured through each of the cameras.

[0011] In a possible implementation manner, in step S3, each of the box images is recognized to obtain the front box number and the rear box number corresponding to the container box as the box number information.

[0012] In a possible implementation manner, in step S3, the recognition server uses a pre-constructed and locally stored damage recognition model to recognize each of the box images to obtain the corresponding box damage condition, and combines a damage label used to represent whether there is damage or not as the damage recognition result.

[0013] In a possible implementation manner, in step S4, when the damage label in the damage recognition result represents that there is damage, the staff at the gate is prompted to perform manual intervention; or when the damage label in the damage recognition result represents that there is no damage, the staff at the gate is prompted to release.

[0014] In a possible implementation manner, after step S4 is executed, it further includes controlling each camera to reset. Description of the Drawings

[0015] Figure 1 It is a flowchart of the steps of the present invention. Detailed Embodiments

[0016] First of all, those skilled in the art should understand that these embodiments are only used to explain the technical principles of the embodiments of the present application, and are not intended to limit the protection scope of the embodiments of the present application. Those skilled in the art can adjust them as needed to adapt to specific application scenarios.

[0017] The following further describes the present application in detail with reference to the drawings and specific embodiments.

[0018] See Figure 1 , the embodiments of the present application disclose an intelligent inspection method for the damage of port container boxes, including: Step S1, determining whether there is a container transport vehicle within the designated positions in the fields of view of each camera: If so, proceed to step S2; If not, return to step S1; Step S2, controlling each camera to perform focal length zooming to take pictures of the container transport vehicle from different directions to obtain box images of different sides of the container box; Step S3, recognizing each box image to obtain box number information, and transmitting each box image to the recognition server for box damage recognition to obtain the corresponding damage recognition result; Step S4, linking the box number information and the damage recognition result with the tally platform to prompt the staff at the gate to perform manual intervention or release.

[0019] The present invention uses a camera in combination with pattern recognition, subverting the entire manual inspection process for damaged containers in traditional docks and completing the entire process of automatically and intelligently inspecting damaged containers, which is as follows: 1. After the container transport vehicle appears at the designated position in the camera's field of view, the box images of 5 sides of the current container are taken respectively through the automatic scanning and focal length scaling of 4 to 6 cameras; 2. The front box number and rear box number of the container are recognized from the box images as box information; 3. The captured box images are sent to the recognition server through the interface; 4. The recognition server recognizes whether the 5 sides of the container are damaged according to the received box images to obtain the corresponding damage recognition results; 5. The box number information, damage recognition results of the recognized damaged container and the tallying platform are linked to prompt the tally clerks for manual intervention; 6. The damaged container is sent to the terminal system and the tallying platform after passing the tally review; 7. The damaged container is linked to the intelligent gate system to prompt the gate personnel for manual intervention; 8. After completing a series of actions of capturing, the camera returns to the original preset position.

[0020] The technical solution of this application is the only method independently proposed by the port party. Only 4 to 6 cameras are required to take pictures of 5 sides of the container. Secondly, it ensures the accuracy of damage recognition without affecting the box number recognition. At the same time, without changing the current operation process and operation efficiency of the container terminal, the inspection and damage operation is integrated, and it is docked and linked with the intelligent tallying and intelligent gate systems. Compared with the traditional manual container inspection and damage means, the number of cameras involved in this application is less, so the cost of completing the inspection and damage operation is lower. Secondly, compared with other solutions that require more cameras to be linked, this application will not reduce the operation efficiency. At the same time, it also ensures safety and reduces the maintenance cost of equipment. Finally, it effectively solves the problem of sharing cameras for damage inspection and box number recognition.

[0021] Generally, the detailed information of the container will be recorded on the outside of the container, such as the box number, container size, box type, container function, loading and unloading status, special box mark, trailer number and trailer box position number, etc. Among them, the box number is a code used to represent information such as the container owner, registration number and check code.

[0022] In a specific business process, after the ship unloading starts, it is divided into two operation lines to record the operation cycles of sea - side locking and operation start respectively. After the sea - side locking, the spreader crossbeam operation starts, followed by the spreader lowering and container identification operations. After the operations are completed, a damage inspection instruction is triggered. At this time, it is divided into two operation lines to record the spreader unlocking and operation cycle respectively. After the spreader is unlocked, it means the operation is completed. After the operation is completed, the damage result is sent and recorded, then the preliminary damage inspection mark is made, and finally the manual verification and confirmation are carried out.

[0023] It should be noted that multiple methods can be used to collect the container number information of the container body. For example, in a specific implementation manner of this application, the container number information of the container body to be tallied is the information obtained by performing image recognition on the container body image captured by an image acquisition device, i.e., a camera, deployed at the port operation site. Specifically, the image acquisition device can be deployed at the gate of the port operation site. After the image acquisition device captures the container body image of the container, the existing image recognition algorithm in the art can be used to recognize the container body image. For example, the OCR (Optical Character Recognition) algorithm can be used to recognize the container body image. Of course, the specific image recognition algorithm adopted in the embodiments of this application does not need to be limited, and any possible implementation manner can be applied to this application.

[0024] In the actual operation process of this application, the qualification of the container body image can also be recognized and judged. Specifically: obtain the target confidence value of the container body image; among them, the target confidence value is a value calculated according to the clarity of the container body image and used to indicate the credibility of the container number information; the above - mentioned target confidence value of the container body image is calculated according to the clarity of the container body image. It can be seen that the target confidence value can reflect the clarity of the collected container body image. It can be understood that the higher the clarity of the container body image, the higher the accuracy of the container number information obtained when using the image recognition algorithm to recognize the container body image, and correspondingly, the higher the credibility of the container number information. On the contrary, the lower the clarity of the container body image, the lower the accuracy of the container number information obtained when using the image recognition algorithm to recognize the container body image, and correspondingly, the lower the credibility of the container number information; judge whether the target confidence value is greater than a preset threshold. If it is, it is determined that the container body image is qualified; if not, it is determined that the container body image is unqualified, and the camera is controlled to re - capture the container body image.

[0025] In step S1, obtain the real-time images captured by each camera, and determine whether there is a container transporter in the specified field of view position corresponding to each real-time image. If so, proceed to step S2; if not, return to step S1. Among them, for each camera in different orientations, the real-time image captured by each camera is divided into separate regions, the box-limiting region is set, and the box-limiting region after setting for each real-time image is used as the specified field of view position.

[0026] In step S3, the recognition server uses the pre-constructed and locally stored damage recognition model to recognize each box image to obtain the corresponding box damage situation, and combines the damage label used to represent the presence or absence of damage as the damage recognition result. When the damage label in the damage recognition result represents the presence of damage, prompt the staff at the gate to perform manual intervention; or when the damage label in the damage recognition result represents the absence of damage, prompt the staff at the gate to release the container.

[0027] In the embodiment of the present application, the YOLOv8 model is used for damage recognition of container boxes. The YOLOv8 model is divided into five versions: n, s, m, l, and x. The model volume increases successively, and the accuracy also improves successively. It mainly consists of Input, Backbone, Neck, and Head. The input end preprocesses the box image. The backbone extracts the features of the input box image through Conv, C2f, and SPPF (spatial pyramid pooling fast) modules. The C2f module draws on the design ideas of C3 and ELAN (efficient layer aggregation network) modules, adds parallel cascading operations to obtain rich gradient flow information. The SPPF module combines feature maps through pooling with different kernel sizes and inputs the result into the Neck. The Neck adopts the PANet (path aggregation network) structure, and performs multi-level fusion of target features at different scales through upsampling and downsampling paths to promote information transmission. The output end Head decouples the classification and prediction processes to obtain the category and position information of the target. The loss calculation includes classification loss BCE Loss, regression loss CIOU Loss, and Distribution Focal Loss.

[0028] In summary, the present application proposes a complete method for intelligent inspection of container box damage at a container terminal, which solves the problem of ensuring the safety of containerized goods at traditional terminals, and can more efficiently, at lower cost, and accurately conduct intelligent inspection of the integrity of container boxes, thereby bringing efficiency improvement to the safety operation guarantee and safety operation inspection of the entire port industry.

[0029] In the description of the present application, the descriptions referring to terms such as "one embodiment", "some embodiments", "in this embodiment", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0030] As described above, the above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for intelligent inspection of damaged container bodies at a port, characterized in that: The tallying platform is applied to the terminal tallying system, the terminal tallying system also includes an identification server, and a plurality of cameras in different directions are pre-installed at the gate of the terminal scene. The port container box damage intelligent inspection method includes the following steps: Step S1, determining whether there is a container transport vehicle within the specified position of the field of view of each camera: If yes, go to step S2; If not, return to step S1; Step S2, controlling each of the cameras to zoom in and out to shoot the container transport vehicle from different directions to obtain images of different surfaces of the container body; Step S3, identifying each of the box images to obtain box number information, and transmitting each of the box images to the recognition server to perform box damage recognition to obtain a corresponding damage recognition result; Step S4, linking the container number information and the damage identification result with the tally platform to prompt the staff at the gate to perform manual intervention or release.

2. The intelligent inspection method for port container damage according to claim 1 is characterized in that: In step S1, the real-time images captured by each camera are obtained, and it is determined whether the container transport vehicle exists in the corresponding designated position of the field of view in each real-time image: If yes, go to step S2; If not, return to step S1.

3. The intelligent inspection method for port container damage according to claim 2 is characterized in that: In the step S1, for each camera in different orientations, the real-time image captured by each camera is divided into separate areas, a box-limited area is set, and the box-limited area after each real-time image is set is used as the field of view designated position.

4. The method for intelligent inspection of damaged port container bodies according to claim 1 is characterized in that: If the number of the cameras is 4-6, then in step S2, the images of the container body on the five surfaces of the container body are obtained by photographing each camera.

5. The method for intelligent inspection of damaged container bodies at a port according to claim 1, characterized in that: In the step S3, each of the box images is identified to obtain the front box number and the rear box number corresponding to the container box as the box number information.

6. The method for intelligent inspection of damaged container bodies at a port according to claim 1, characterized in that: In step S3, the recognition server uses a damage recognition model pre-built and stored locally to recognize each of the box images to obtain the corresponding box damage status, and combines the damage label used to characterize whether damage exists as the damage recognition result.

7. The intelligent inspection method for port container damage according to claim 6 is characterized in that: In step S4, when the damaged label in the damage identification result indicates damage, the staff at the gate is prompted to perform manual intervention; or when the damaged label in the damage identification result indicates no damage, the staff at the gate is prompted to release the vehicle.

8. The method for intelligent inspection of damaged container bodies at a port according to claim 1, characterized in that: After executing step S4, the method further includes controlling each of the cameras to reset.

Images