Gantry crane instruction execution autonomous confirmation method based on external camera installation

Abstract

The invention discloses a gantry crane instruction execution autonomous confirmation method based on external camera installation, and relates to the technical field of mechanical automation, and the method mainly comprises the following steps: carrying out execution confirmation of a gantry crane displacement instruction under station real-time determination through combination of preliminary operation block positioning under satellite positioning and a lateral camera; performing execution confirmation of a first business instruction under target container truck number matching and a container truck alignment guide instruction under parking space matching in sequence through a lateral camera; performing execution confirmation of a second business instruction under matching of the container number of the target container through the corresponding camera; executing confirmation of checking the hoisting instruction of the lower hoisting tool according to the operation action requirement through a top camera; and feature matching and execution confirmation of a lifting appliance lifting and unloading instruction under operation action requirement checking are carried out through a top camera. According to the invention, accurate sensing of gantry crane operation and automatic confirmation of instructions are realized, dependence on an internal PLC system of the gantry crane is avoided, and hardware modification cost and complexity are reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical automation, in particular to a gantry crane instruction execution autonomous confirmation method based on external camera installation. BACKGROUND

[0002] In the port container yard operation, the gantry crane (RTG) is the core handling equipment, and its operation efficiency and accuracy directly affect the operation efficiency of the entire terminal. The traditional gantry crane operation process highly depends on the naked eye observation and manual operation of the operator driver, for example, the driver needs to judge whether the work station after the gantry crane moves is accurate, whether the truck is aligned, whether the box number of the container to be operated is correct, whether the lock of the spreader and the container is in place, etc., and then manually confirms the completion of the instruction execution on the terminal. This manual operation mode not only causes great physical and mental burden to the driver, but also has problems such as subjective judgment error, confirmation action lag or advance, which leads to the asynchronization of the instruction execution state and the system plan state, becoming the bottleneck of improving the precision and real-time of the yard scheduling. In order to improve the automation level, the industry has gradually developed the remote control (remote control) gantry crane technology, which realizes the state perception and automatic control by installing sensors on the gantry crane and modifying the PLC (Programmable Logic Controller) system. However, this modification scheme is high in cost and long in cycle, and the equipment needs to be shut down during the modification, which seriously affects the production operation. The more realistic problem is that the manually operated gantry crane and the remote control gantry crane will coexist in the same operation site for a long time. Therefore, the market urgently needs a solution that can low-cost and quickly modify the existing manual gantry crane and can simultaneously compatible with the future remote control system for automatic state perception and instruction confirmation. Some existing technologies try to rely on single satellite positioning for rough positioning, or try to install a large number of expensive laser radars, encoders and other sensors to obtain operation information, which all have obvious defects: the former has insufficient accuracy and cannot meet the fine operation requirements such as alignment and box number identification; the latter makes the system complex, reduces the reliability, greatly increases the modification and maintenance costs, and is difficult to get rid of the deep dependence on the gantry crane PLC system, which has poor engineering landing performance. SUMMARY

[0003] In order to realize a low-cost, simple-to-deploy solution that neither relies on complex sensors nor depends on the gantry crane internal PLC system, and yet can fully, accurately and automatically complete the gantry crane operation whole-process state perception and instruction confirmation, the present application proposes a gantry crane instruction execution autonomous confirmation method based on external camera installation, at least two lateral cameras facing the operation lane and arranged at intervals are arranged on the left and right support arms of the gantry crane, and a top camera horizontally downward is arranged at the bottom of the gantry crane trolley, which includes the following steps: S1: Through the preliminary work block positioning under satellite positioning, combined with the real-time acquisition of ground signs by lateral cameras, the gantry crane is determined in real time, and the execution of the gantry crane shifting instruction is confirmed; S2: After the execution of the gantry crane shifting instruction is confirmed, the truck number and the truck position are recognized by the lateral camera, and the execution of the first business instruction under the target truck number matching and the truck alignment guide instruction under the position matching are confirmed in turn; S3: After the execution of the first business instruction is confirmed, the container number is recognized by the corresponding camera during the horizontal movement of the gantry crane and / or the spreader according to the work requirements, and the execution of the second business instruction under the target container number matching is confirmed; S4: After the execution of the second business instruction is confirmed, the relative height information between the spreader and the target container is acquired by the top camera, and the execution of the spreader lifting instruction under the work action requirement check is confirmed; S5: After the execution of the spreader lifting instruction is confirmed, the feature information of the target stacking position and the relative height information between the target container and the target stacking position are acquired by the top camera according to the work requirements, and the execution of the spreader lifting and unloading instruction under the feature matching and work action requirement check is confirmed.

[0004] Further, the gantry crane work area includes a plurality of work blocks, and each work block includes sequentially coded ground signs.

[0005] Further, in the S1 step, when the gantry crane needs to turn on the road, it further includes the steps of: S11: The turning angle of the gantry crane tire is recognized by the lateral camera, and the execution of the gantry crane turning instruction under the turning angle check is confirmed.

[0006] Further, the method of recognizing the truck number and the container number is specifically: based on the proportion of the recognized numbers meeting the pre-set confidence, after meeting the proportion requirement, the corresponding truck number or container number is called from the database.

[0007] Further, in the S2 step, the position matching is specifically: a virtual vertical line representing an ideal alignment area is pre-set in the image captured by the lateral camera; by recognizing the relative position of the truck frame or the container and the virtual vertical line, the alignment deviation value of the truck is calculated.

[0008] Further, the S2 step further includes the step of: feeding back the alignment deviation value to the display screen in the truck cab in real time to guide the driver to complete the alignment guide in a graphical way.

[0009] Further, the relative height information is specifically acquired by the following steps: A set of virtual horizontal lines of preset heights are preset in the image captured by the top camera; by recognizing the relative position or intersection state of the specific part of the spreader or container with the virtual horizontal lines, it is determined whether the spreader and the container reach the relative height required by the operation action.

[0010] Further, the operation requirement includes lifting the container from the stack to the truck, and lifting the container from the truck to the stack.

[0011] Further, in the S5 step, when the operation requirement is to lift the container from the stack to the truck, the feature information is the contour of the truck base, and when the operation requirement is to lift the container from the truck to the stack, the feature information is the contour of the top frame of the container.

[0012] Compared with the prior art, the present application has at least the following beneficial effects: (1) The gantry crane instruction execution autonomous confirmation method based on the external camera added by the present application realizes accurate perception of the whole life cycle state of the gantry crane operation and automatic confirmation of the instruction by adding five cameras at the key positions of the gantry crane and assisting with machine vision algorithms, which breaks the dependence on the internal PLC system of the gantry crane and reduces the hardware modification cost and engineering complexity; (2) By fusing satellite preliminary positioning and recognition of ground marks by lateral cameras, the accuracy of the position determination of the gantry crane is ensured; (3) By presetting virtual vertical lines, accurate calculation and real-time visual guidance of the deviation of the truck alignment are realized, which significantly improves the alignment efficiency and safety; (4) The virtual horizontal lines in the image are used to replace the traditional physical height sensor, which successfully realizes reliable judgment of the operation layer height and the relative position of the spreader and the container, and completes automatic checking of the lifting and unloading instructions. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 The steps of a gantry crane instruction execution autonomous confirmation method based on an external camera added. DETAILED DESCRIPTION

[0014] The following is a specific embodiment of the present application and further describes the technical solution of the present application in conjunction with the drawings, but the present application is not limited to these embodiments.

[0015] The prior art relies on the driver's artificial judgment or a deep reformed PLC and sensor scheme, which is difficult to balance between economy, reliability and implementation convenience, and cannot meet the urgent needs of current wharfs for a large number of manual gantry cranes to be upgraded in a low-cost and efficient automatic manner. In order to overcome the above defects, the present application provides a brand-new technical idea, that is, through a small number of external cameras with optimized layout and a set of software algorithms deeply integrating artificial intelligence and machine vision, the eyes and brain of human drivers are simulated and replaced to realize autonomous perception and judgment of the whole operation process. Figure 1 As shown in FIG. 1, the present application proposes a gantry crane instruction execution autonomous confirmation method based on external camera installation, at least two lateral cameras facing the operation lane and arranged at intervals are arranged on the left and right support arms of the gantry crane, and a top camera horizontally downward is arranged at the bottom of the gantry crane trolley, which comprises the following steps: S1: through the preliminary operation block positioning under satellite positioning and the real-time acquisition of ground marks by the lateral camera, the execution confirmation of the gantry crane displacement instruction is determined in real time; S2: after the execution confirmation of the gantry crane displacement instruction, the container truck number and the container truck position are recognized by the lateral camera, and the execution confirmation of the first business instruction under the target container truck number matching and the container positioning guide instruction under the position matching are sequentially performed; S3: after the execution confirmation of the first business instruction, the container box number is recognized by the corresponding camera during the horizontal movement of the gantry crane and / or the spreader according to the operation requirement, and the execution confirmation of the second business instruction under the target container box number matching is performed; S4: after the execution confirmation of the second business instruction, the relative height information between the spreader and the target container is acquired by the top camera, and the execution confirmation of the spreader hoisting instruction under the operation action requirement checking is performed; S5: after the execution confirmation of the spreader hoisting instruction, the feature information of the target stacking position and the relative height information between the target container and the target stacking position are recognized by the top camera according to the operation requirement, and the execution confirmation of the spreader hoisting and unloading instruction under the feature matching and operation action requirement checking is performed.

[0016] In the specific implementation process, the autonomous confirmation system first acquires the approximate longitude and latitude coordinates of the gantry crane through the satellite positioning module (such as GNSS receiver) installed on the gantry crane, and through comparison with the electronic map of the yard where the gantry crane is planned to operate, it can be preliminarily determined that it is near which large operation block. However, due to the several-meter error of satellite signal itself and the multi-path interference caused by container stacking, equipment, etc. in the yard, only satellite positioning cannot meet the centimeter to decimeter level accurate parking requirement required by the gantry crane operation.

[0017] To solve this problem, at least two lateral cameras are arranged on the left and right support legs of the gantry crane, facing the work lane side. These cameras continuously take pictures of the ground during the movement of the gantry crane. In practice, the ground of the work area is pre-paved with sequentially coded ground markings, which are accurately installed at specific positions of each bay according to the work process requirements. When the gantry crane moves, the lateral cameras continuously capture the ground images within their field of view and extract the ground markings from the images.

[0018] At this time, the autonomous confirmation system fuses and verifies the rough block information provided by satellite positioning with the accurate identification code information recognized by visual recognition. For example, satellite positioning determines that the gantry crane is near the "A05" block, and the camera recognizes the ground marking code as "A05B02" (representing the 2nd bay of the A05 block). The system can accurately determine the exact work position of the gantry crane at this time. Only when the real-time work position after moving matches the target work position information issued by the scheduling system, the system will automatically generate and execute the "gantry crane moving instruction execution confirmation" signal, and report the status to the central control system in real time, indicating that the moving operation is completed. This step greatly reduces the driver's manual confirmation burden and lays the foundation for subsequent series of fine work instruction automatic confirmation. If the gantry crane needs to switch between work areas and pass through the turning lane, it also needs to further monitor the tire turning angle through the lateral camera to ensure that the turning action is executed safely according to the preset path and is confirmed after completion.

[0019] After confirming that the gantry crane has accurately moved to the target work bay, the lateral high-definition camera deployed on the support arm of the gantry crane captures the image of the tractor vehicle moving or parked in the work lane. First, the autonomous confirmation system calls the tractor number recognition algorithm to process the captured image. This algorithm is based on a deep learning model and can resist interference such as light changes, partial occlusion, and vehicle body dirt, accurately positioning and identifying the license plate number on the trailer or the work number on the box. To improve the accuracy and reliability of identification, the system does not require 100% perfect identification of all characters, but uses an intelligent matching mechanism based on pre-set confidence: when a high enough proportion (e.g. more than 80%) of characters in the identified character sequence reach a very high confidence, the system determines that this identification is valid and immediately compares the result with the target tractor number in the work instruction issued by the scheduling system. Only when they completely match, the system will generate a "first business instruction (usually'start working on this tractor') execution confirmation" signal, thereby ensuring the absolute correctness of the work object and avoiding the risk of misoperation from the source.

[0020] After confirming "for whom to work", the autonomous system automatically proceeds to the judgment of "parking space alignment", which is the core of ensuring work safety and efficiency. The traditional scheme needs to rely on sensors installed on the truck or expensive laser scanning equipment, while the invention adopts "software-defined" virtual detection technology. In the video stream of the camera, the system predefines one or more invisible "virtual vertical lines" or a "virtual area" in the image according to the ideal parking position, which represents the best reference position of the truck trailer or container when it is ideally aligned. The computer vision algorithm analyzes the image in real time, calculates the pixel deviation between the truck trailer edge, corner or container contour and the virtual reference line, and then converts the pixel deviation into an accurate actual distance deviation value (such as centimeter level) according to the pre-calibrated camera parameters. This calculation process is continuous and real-time, and if the system judges that the current alignment deviation is within the allowable safety tolerance range, it will automatically generate a "truck alignment guidance instruction execution confirmation" signal. At the same time, in order to further improve the operation experience and cooperation efficiency of the ground driver, the system can also send the calculated alignment deviation value to the intelligent display screen equipped in the truck cab through wireless communication, to guide the driver to make fine adjustments with intuitive arrow icons and numerical prompts, forming a closed-loop guidance system until perfect alignment is achieved. Thus, the full-automatic judgment from vehicle identity confirmation to work position confirmation is completed.

[0021] During the parking space alignment process, the gantry crane simultaneously executes the second service instruction, and uses the time window of the spreader moving above the truck or stack to dynamically activate the corresponding recognition camera. If the target container is located on the truck trailer, the continuously working lateral camera is responsible for capturing the container number image on the side of the container; if the target container is located in the stack, the gantry crane adjusts the pose of the trolley or car to enable the top camera installed at the bottom of the hoist trolley to capture the container number at the best angle, and to identify the container number at the center of the field of view in real time until the multiple identified container numbers are the same, and the spreader has a downward trend (visual), the container number and the work instruction information are consistent, and the instruction is confirmed. This dynamic identification strategy minimizes the pure waiting time and perfectly embeds the identification process into the inherent movement process of the equipment, greatly improving the work efficiency.

[0022] In this process, the box number recognition algorithm built-in the autonomous confirmation system is trained through massive port scene images, so as to recognize and adapt to box number font, format, wear, tilt, partial occlusion and complex light changes (such as reflection, shadow). The algorithm first accurately locates the box number area in the image, and then performs character segmentation and recognition. Similar to the vehicle number recognition, the high confidence matching principle is also adopted here: the system does not require perfect recognition of all characters at one time, but calculates the proportion of characters that have been successfully recognized and have very high confidence in the complete box number, and once the proportion exceeds the preset threshold (such as 80%), it is considered as effective recognition. Subsequently, the system immediately compares the recognized box number character sequence with the target box number specified in the dispatching instruction. This comparison process also has certain fault tolerance intelligence and can handle the case of individual character identification ambiguity caused by dirt. Only when the container number in the current field of view is highly consistent with the target box number, the system will automatically generate a "second business instruction (usually 'confirm operation this container') confirmation" signal. The reporting of this confirmation signal means that the identity of the target container has been "locked" by the system, and then the spreader can perform precise operations such as lowering and lifting or releasing and unloading on the specified container, thereby eliminating major operation accidents caused by driver's visual misjudgment or manual input errors, and laying a foundation for the reliability of the entire automation process.

[0023] During the lifting process, unlike the traditional scheme which relies on laser or ultrasonic ranging sensors installed on the spreader, the present application adopts a "virtual horizontal line" ranging method based on vision. Specifically, during the initialization phase of the autonomous confirmation system, the top camera is first accurately calibrated to establish a mapping relationship between image pixel coordinates and actual physical space. Then, according to different operating conditions (such as 20-foot containers or 40-foot containers), spreader types and safety operating distance requirements, one or more "virtual horizontal lines" representing different absolute heights or relative heights are defined in the real-time image captured by the camera. These lines do not exist in the physical world, but are only reference benchmarks embedded in the image processing algorithm.

[0024] When the spreader starts to descend and approaches the target container, the top camera continuously captures images. The vision algorithm analyzes the video stream in real time, dynamically tracking the specific feature parts of the spreader (such as the spinlock head) or the contour edge of the container top plate. The system accurately determines the current relative height by calculating the relative positional relationship between these feature parts in the image and the preset virtual horizontal line representing the "ideal locking height". For example, when the algorithm identifies that the image of the spreader's spinlock head is about to contact or coincide with the virtual horizontal line, it determines that the spreader has descended to the precise height at which the locking action can be safely performed. This judgment process is continuous and real-time, and the system can predict in advance and generate a confirmation signal at the moment when the height reaches the set threshold. Once it is confirmed that the current height meets the work requirements, the system will automatically generate a "spreader lifting instruction execution confirmation" signal. This method not only realizes non-contact accurate measurement of height, avoiding problems such as sensor installation, wiring maintenance, and mechanical collision damage, but also gives the system flexibility and adaptability through pure visual software algorithms - simply by modifying the position parameters of the virtual line, different types of spreaders or containers can be easily adapted without any hardware changes, perfectly embodying the core advantage of "software definition" and providing an economical and reliable technical path for the automatic upgrade of gantry cranes.

[0025] When the spreader is lifting the container to the target stacking position, it is again realized by recognizing the feature information of the target stacking position through the top camera located at the bottom of the trolley. If the work instruction is to load the container onto the truck, the algorithm will focus on identifying the contour of the twistlock frame or the structural features of the trailer flatbed to confirm that the spreader is directly above the trailer, not in other incorrect positions. If the work instruction is to unload the container to the stack, the target feature becomes the corner profile or top plate boundary of the topmost container of the lower stack, and the system matches these features to ensure that the container to be unloaded is in the correct alignment state with the containers on the stack, preventing the risk of collapse or collision.

[0026] After the feature matching is successfully performed and the planar position is confirmed to be correct, the system starts the most critical height determination process, which also depends on the "virtual horizontal line" defined in the image by the software in advance. However, the physical height represented by the virtual line is set as the final height of the safe landing of the container. When the spreader starts to descend, the algorithm analyzes the image in real time, tracks the relative position between the virtual line representing the "safe landing height" and the bottom of the hoisted container (for placing on the dolly) or the top frame of the target stacked container (for placing on the stack). By calculating the pixel distance change of the two in the image, the system can accurately perceive the change of the relative height. When the image of the container bottom and the dolly frame, or the image of the bottom of the container to be placed and the top frame of the top container of the stack, is about to coincide with the virtual safe line, the system accurately determines that the height at this moment is the optimal release height, and at this moment, the system automatically generates a "spreader lifting and unloading instruction execution confirmation" signal. This whole process does not require the driver to make any visual estimation and manual confirmation of the height or position, and is completely completed automatically by the autonomous confirmation system, which not only greatly improves the safety and accuracy of the operation, eliminates the impact caused by releasing too high or the container bumping caused by insufficient release height, but also improves the operation efficiency.

[0027] In summary, the gantry crane instruction execution autonomous confirmation method based on the external camera proposed by the application realizes accurate perception of the whole life cycle state of the gantry crane operation and automatic confirmation of the instruction by installing five cameras at key positions of the gantry crane and assisting with machine vision algorithms, breaks the dependence on the internal PLC system of the gantry crane, and reduces the hardware modification cost and engineering complexity.

[0028] And by fusing the satellite preliminary positioning and the identification of the ground marks by the lateral camera, the accuracy of the position determination of the gantry displacement is ensured; by the preset virtual vertical line technology, the accurate calculation and real-time visual guidance of the deviation of the truck alignment are realized, and the alignment efficiency and safety are significantly improved; the virtual horizontal line in the image is used to replace the traditional physical height sensor, the reliable judgment of the operation layer height and the relative position of the spreader and the container is successfully realized, and the automatic checking of the lifting and unloading instruction is completed.

[0029] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the application are only used to explain the relative position relationship, motion condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directional indications also change accordingly.

[0030] In addition, in the present application, the description such as "first", "second", "one" and the like are only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the technical features or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise explicitly specified and limited.

[0031] In the present application, unless otherwise explicitly specified and limited, the terms "connection", "fixing" and the like should be understood broadly, for example, "fixing" can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be directly connected, or indirectly connected through intermediate medium, can be the internal communication of two elements or the interaction relationship between two elements, unless otherwise explicitly limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0032] In addition, the technical solutions of various embodiments of the present application can be combined with each other, but it must be based on the realization of the ordinary skilled in the art, when the combination of technical solutions appears contradictory or cannot be realized, it should be considered that the combination of technical solutions does not exist, nor within the scope of protection required by the present application.

Claims

1. A gantry crane instruction execution autonomous confirmation method based on an externally mounted camera, characterized in that, The gantry crane is provided with at least two lateral cameras facing the operation lane and arranged at intervals on the left and right support arms of the gantry crane, and a top camera horizontally downward arranged at the bottom of the gantry crane trolley, and comprises the following steps: S1: through the preliminary operation block positioning under satellite positioning and the real-time acquisition of ground marks by the lateral cameras, the execution confirmation of the gantry crane displacement instruction is performed under the real-time determination of the working position of the gantry crane; S2: after the execution confirmation of the gantry crane displacement instruction, the truck number and the truck position are recognized by the lateral cameras, and the execution confirmation of the first business instruction under the target truck number matching and the truck alignment guide instruction under the truck position matching are performed in sequence; S3: after the execution confirmation of the first business instruction, the container number is recognized by the corresponding camera during the horizontal movement of the gantry crane and / or the spreader according to the operation requirement, and the execution confirmation of the second business instruction under the target container number matching is performed; S4: after the execution confirmation of the second business instruction, the relative height information between the spreader and the target container is acquired by the top camera, and the execution confirmation of the spreader hoisting instruction under the operation action requirement checking is performed; S5: after the execution confirmation of the spreader hoisting instruction, the feature information of the target stacking position and the relative height information between the target container and the target stacking position are recognized by the top camera according to the operation requirement, and the execution confirmation of the spreader hoisting and unloading instruction under the feature matching and the operation action requirement checking is performed.

2. The gantry crane instruction execution autonomous confirmation method based on an externally mounted camera according to claim 1, characterized in that, The gantry crane operation area comprises a plurality of operation blocks, and each operation block comprises sequentially coded ground marks.

3. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 1, characterized in that, In the S1 step, when the gantry crane needs to pass the turning lane, the following steps are further included: The turning angle of the gantry crane tire is recognized by the lateral camera, and the execution confirmation of the gantry crane turning lane instruction under the turning angle checking is performed.

4. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 1, characterized in that, The method for recognizing the truck number and the container number is specifically: based on the proportion of the recognized numbers meeting the pre-set confidence, the corresponding truck number or container number is called from the database after meeting the proportion requirement.

5. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 1, characterized in that, In the S2 step, the truck position matching is specifically: a virtual vertical line representing an ideal alignment area is pre-set in the image captured by the lateral camera; by recognizing the relative position of the truck frame or the container and the virtual vertical line, the alignment deviation value of the truck is calculated.

6. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 5, characterized in that, In the S2 step, the following step is further included: the alignment deviation value is fed back to the display screen in the truck cab in real time, so as to guide the driver to complete the alignment guide in a graphical manner.

7. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 1, characterized in that, The relative height information is specifically acquired by the following steps: A group of virtual horizontal lines of pre-set heights are pre-set in the image captured by the top camera; by recognizing the relative position or intersection state of the specific part of the spreader or the container and the virtual horizontal line, it is judged whether the spreader and the container reach the relative height required by the operation action.

8. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 1, characterized in that, The operation requirement includes hoisting the container from the stack to the truck and hoisting the container from the truck to the stack.

9. The gantry crane command execution autonomous confirmation method based on an externally mounted camera according to claim 8, characterized in that, In the S5 step, when the operation requirement is to hoist the container from the stack to the truck, the feature information is the truck base contour, and when the operation requirement is to hoist the container from the truck to the stack, the feature information is the container top frame contour.

Images