Intelligent anti-collision warning method and system for ship loader

Abstract

The present invention relates to the field of data processing, and more specifically to an intelligent anti-collision warning method and system for ship loading machines. The method comprises: obtaining pre-processed continuous multi-frame point cloud data of the ship loading machine and its surrounding area; calculating the local density of each point cloud data, and setting an initial clustering radius based on the local density; selecting any point cloud data as a target point cloud, obtaining the similarity value of all neighboring point clouds within the initial clustering radius of the target point cloud, and adjusting the initial clustering radius of the target point cloud using the similarity value to obtain an optimal clustering radius; using the optimal clustering radius to cluster and segment the single-frame point cloud data into different clusters, each cluster representing a target, calculating the distance between the target and the chute, and triggering an anti-collision warning based on the distance. The present invention achieves accurate identification and distance measurement of targets around the ship loading machine, improving the accuracy of the anti-collision warning and the safety of the ship loading operation.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more specifically, to an intelligent anti-collision warning method and system for a ship loader. Background Art

[0002] Ship loaders are large-scale machines used at bulk material terminals. Their primary function is to load bulk materials (such as coal, ore, and grain) from the terminal's storage area into the holds of cargo ships. The chute is a key component of the ship loader, located at the terminal end. Its function is to guide materials from the ship loader's conveyor system into the ship's hold. Because the chute needs to be constantly moving to ensure uniform loading, and the ship's draft changes dynamically with the load, the relative position of the chute and hold also constantly fluctuates. To prevent collisions between the chute and the hold and ensure the safety and efficiency of loading operations, a collision warning system is essential.

[0003] Currently, anti-collision warning systems typically use radar to acquire three-dimensional point cloud data in the ship loader's working environment. The system segments this point cloud data to identify target obstacles (such as bulkheads), calculates the distance between these target obstacles and the chute, and issues anti-collision warnings based on the distance.

[0004] However, since the point cloud data obtained by radar often contains a large amount of noise and interference information, such as changes in ambient light, differences in the reflection characteristics of objects, errors in radar equipment, etc., these noise and interference information will interfere with the accuracy of the point cloud data, making it difficult for the segmentation algorithm to accurately identify obstacles, which in turn causes the anti-collision system to issue unnecessary warning signals, thereby interfering with the normal operation of the ship loader. Summary of the Invention

[0005] In order to solve the above-mentioned technical problem that obstacles cannot be accurately identified, resulting in the anti-collision system issuing unnecessary warning signals and interfering with the normal operation of the ship loader, the present invention provides solutions in the following aspects.

[0006] In a first aspect, an intelligent anti-collision warning method for a ship loader includes:

[0007] Acquire pre-processed continuous multi-frame point cloud data of the ship loader and a certain area around it, wherein the certain area around it includes the chute, the cabin, and the cargo;

[0008] Calculating the local density of each point cloud data, and setting the initial clustering radius according to the local density;

[0009] Select any point cloud data as the target point cloud, obtain all neighboring point clouds within the initial clustering radius of the target point cloud, calculate the similarity value between the target point cloud and each of the neighboring point clouds, use the similarity value to adjust the initial clustering radius of the target point cloud, obtain the optimal clustering radius of the target point cloud, and then obtain the optimal clustering radius of all point cloud data; the similarity value is positively correlated with the speed consistency between the target point cloud and its neighboring point clouds, and is negatively correlated with the distance between the target point cloud and the neighboring point clouds;

[0010] The optimal clustering radius is used to cluster the point cloud data of a single frame into different clusters, each cluster represents a target, and the distance between the target and the chute is calculated. The distance is compared with a preset threshold. If the distance is less than the threshold, an anti-collision warning is triggered.

[0011] The present invention first calculates the initial clustering radius based on a preset clustering radius benchmark value and local density, so that clustering can adapt to areas with different densities. Secondly, the speed consistency and distance of continuous multi-frame point cloud data are tracked, and the initial clustering radius is dynamically adjusted. Even if there are temporarily docked cargoes near the chute (with close spatial distance but large speed difference), they can be segmented into different clusters to avoid misjudgment. Finally, clustering is performed using the optimal clustering radius, which can accurately separate key targets such as ship loaders, chute, and cargo, avoiding deviations in distance calculation caused by incorrect clustering.

[0012] In summary, the present invention improves the accuracy of segmentation and distance detection by dynamically adjusting clustering parameters and combining multimodal information (space and speed), thereby effectively triggering early warnings and improving operational safety.

[0013] Preferably, the point cloud data includes three-dimensional coordinates and reflection intensity of the point cloud data.

[0014] Preferably, the process of obtaining the local density includes:

[0015] Normalize the three-dimensional coordinates and reflection intensity of the point cloud data, and use kernel density estimation to calculate the local density of each point cloud data;

[0016] The Gaussian kernel function is selected and the bandwidth of the kernel density estimate is calculated.

[0017] Kernel density estimation is a non-parametric method that does not rely on prior distribution assumptions of the data and is applicable to a variety of complex and unknown distribution forms. For point cloud data, whose distribution can be very complex, kernel density estimation can flexibly estimate its local density.

[0018] Preferably, the initial cluster radius satisfies the relationship:

[0019] Where, For the The initial clustering radius of point cloud data, is the cluster radius benchmark value, is the global density of all point cloud data, For the The local density of all point cloud data; the global density of all point cloud data is the average of the local densities of all point cloud data.

[0020] In point cloud data, noise and outliers may cause abnormal fluctuations in local density. By introducing global density constraints, the algorithm's tolerance to these local anomalies will be enhanced, thereby improving the robustness of the clustering results.

[0021] Preferably, the similarity satisfies the relationship:

[0022] Where, For the The target point cloud and the The similarity value of the neighborhood point cloud, For the The target point cloud and the The velocity consistency between neighboring point clouds, For the The target point cloud and the The Euclidean distance between neighboring point clouds, is an exponential function with the natural constant e as the base; By calculating the The target point cloud and the The cosine similarity between the velocity vectors of the neighborhood point clouds is obtained.

[0023] By combining velocity consistency and distance information, it is possible to more accurately determine whether two point clouds belong to the same object or are related. Especially in dynamic scenes, velocity information can help distinguish moving targets and avoid mismatching.

[0024] Preferably, the optimal clustering radius satisfies the relationship:

[0025] Where, For the The optimal clustering radius of the target point cloud, For the The initial clustering radius of the target point cloud, For the The target point cloud and the The similarity value of the neighborhood point cloud, is the total number of neighborhood points of the target point cloud, It is an exponential function with the natural constant e as its base.

[0026] By introducing a weighted average of similarity values, the formula can dynamically adjust the initial cluster radius based on the local characteristics of the point cloud. In areas of high similarity, the cluster radius is larger, allowing more points to be included; in areas of low similarity, the cluster radius is smaller, avoiding the inclusion of too many dissimilar points. This improves clustering accuracy and efficiency.

[0027] Preferably, obtaining the optimal clustering radius of all point cloud data further includes:

[0028] Use the optimal clustering radius to segment the point cloud data of a single frame to obtain multiple clusters; select any two consecutive frames of point cloud data and align them;

[0029] For each point cloud data in the first frame, calculate the information entropy of the cluster label of the corresponding point cloud data in the second frame;

[0030] An objective function is established, the sum of all information entropies is input, and the optimal clustering radius benchmark value is output; based on the optimal clustering radius benchmark value, the adjusted optimal clustering radius is obtained.

[0031] Adjusting the cluster radius is an iterative optimization process. The cluster radius baseline provides an initial reference, and the introduction of local features makes the initial cluster radius more closely aligned with the actual data distribution. However, to further improve clustering accuracy and stability, the optimized initial radius needs to be fed back into the baseline adjustment. This iterative optimization can continuously approach the optimal cluster radius, resulting in more accurate clustering results.

[0032] Preferably, the objective function is expressed as:

[0033] Where, is the objective function, The first frame The information entropy of the cluster labels corresponding to the point cloud data in the cluster in the second frame, is the number of clusters in the first frame, is the optimal clustering radius benchmark value.

[0034] Preferably, the clustering is a DBSCAN clustering algorithm.

[0035] In a second aspect, an intelligent anti-collision warning system for a ship loader is provided, comprising: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, any one of the intelligent anti-collision warning methods for a ship loader is implemented.

[0036] The beneficial effects of the present invention are:

[0037] Through point cloud data processing and cluster analysis, the present invention continuously optimizes clustering parameters to achieve real-time monitoring of the ship loader and its surrounding environment (such as the chute, cabin, and cargo), thereby effectively preventing the ship loader from colliding with other objects during operation and improving the safety and efficiency of ship loading operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a method flow chart of steps S1 to S4 in an intelligent anti-collision warning method for a ship loader according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0040] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0041] Reference Figure 1 A ship loader intelligent anti-collision warning method includes steps S1 to S4, which are specifically as follows:

[0042] S1: Acquire pre-processed continuous multi-frame point cloud data of the ship loader and a certain area around it, wherein the certain area around it includes the chute, the cabin, and the cargo.

[0043] In one embodiment, a laser radar device is installed on the chute maintenance platform of the ship loader, and the laser radar is used to collect point cloud data of the area around the ship loader in real time. These areas include the chute, the cabin, and the cargo in the cabin. The collected point cloud data includes the three-dimensional coordinates (i.e., x, y, z position information) and reflection intensity (i.e., the energy intensity returned after the laser beam hits the surface of an object, which can be used to determine the material, color, and other characteristics of the object) of each point.

[0044] To more comprehensively and accurately capture the dynamics of the ship loader and its surrounding chutes, holds, and cargo, the LiDAR equipment must continuously collect data. For example, a frame of point cloud data is acquired at regular intervals (e.g., every 0.1 second), with multiple frames (e.g., 100 frames) collected continuously.

[0045] Specifically, for each frame of data, based on the three-dimensional coordinate information of each point it contains, all points are organized together according to certain rules (such as spatial position relationship or scanning order) to form an ordered point set.

[0046] Since the data obtained by the lidar may contain noise, methods such as mean filtering can be used to remove the noise data.

[0047] S2: Calculate the local density of each point cloud data, and set the initial clustering radius according to the local density.

[0048] During port operations, ship loaders must maintain a safe distance from surrounding equipment and objects to avoid collisions. Clustering and segmenting point cloud data allows real-time monitoring of the distance and positional relationship between the ship loader and surrounding objects, identifying potential collision risks and implementing appropriate avoidance measures to ensure safe operations.

[0049] The local density of points is a very important factor because it can reflect the spatial distribution characteristics of point cloud data. In high-density areas, the distances between points are close, so a smaller cluster radius is sufficient to include enough points to form a meaningful cluster. Conversely, in low-density areas, the distances between points are greater, and a larger cluster radius is required to include enough points.

[0050] Therefore, the local density of point cloud data needs to be considered when calculating the clustering radius of point cloud data.

[0051] In one embodiment, the general steps for calculating the local density of point cloud data are as follows:

[0052] First, the four-dimensional data (x, y, z, and reflection intensity) of each point cloud data is obtained, and then the data of these dimensions are further normalized and scaled to the same scale.

[0053] Then, the local density of each point cloud data is calculated by kernel density estimation, where a Gaussian kernel function is selected and the bandwidth of the kernel density estimation is calculated using the Silverman rule.

[0054] It should be noted that kernel density estimation is an existing technology and will not be described in detail here.

[0055] Next, according to the local density of each point cloud data, an initial clustering radius is set for it, which satisfies the relationship:

[0056]

[0057] Where, For the The initial clustering radius of point cloud data, is the cluster radius benchmark value, is the global density of all point cloud data, For the The global density of all point cloud data is the mean of the local densities of all point cloud data. The cluster radius benchmark value is used to ensure that the initial cluster radius is not too large or too small. Here, the cluster radius benchmark value is set to 0.01.

[0058] Among them, if the If the local density of a point cloud data is high, it means that there are many points around it, so the initial clustering radius should be reduced to ensure that the high-density area is over-divided into multiple small clusters; on the contrary, if the If the local density of a point cloud data is low, it means that there are few points around it, so the initial clustering radius should be expanded to ensure that clusters in low-density areas can be detected.

[0059] In another embodiment, the initial cluster radius also satisfies the relationship:

[0060]

[0061] Where, For the The initial clustering radius of point cloud data, is the cluster radius benchmark value, is the global density of all point cloud data, For the The local density of point cloud data, For the The weight of each point cloud data point (based on actual conditions, give higher weight to data that requires special attention during the clustering process). The global density of all point cloud data is the average of the local density of all point cloud data. The cluster radius baseline value is used to ensure that the initial cluster radius is not too large or too small. Here, the cluster radius baseline value is set to 0.01.

[0062] S3: Select any point cloud data as the target point cloud, obtain all neighboring point clouds within the initial clustering radius of the target point cloud, calculate the similarity value between the target point cloud and each of the neighboring point clouds, and use the similarity value to adjust the initial clustering radius of the target point cloud to obtain the optimal clustering radius of the target point cloud, and then obtain the optimal clustering radius of all point cloud data.

[0063] When the distance between two point cloud data is slightly larger than the initial clustering radius, but they are actually very similar in other features, if the initial clustering radius is not adjusted, the two point cloud data may be divided into different clusters. Therefore, it is necessary to improve the accuracy and adaptability of clustering through continuous adjustment and optimization based on the initial clustering radius.

[0064] In one embodiment, the radial velocity of the point cloud data is measured by a laser radar, that is, the velocity component of the point cloud along the radar beam direction. The radial velocity of each point cloud data needs to be decomposed into three-dimensional coordinates, and the radial velocity is further multiplied by the unit vector of the corresponding three-dimensional coordinate of the point cloud data to obtain the three-dimensional velocity sequence of each point cloud data ( 、 、 ).

[0065] Then, any point cloud data is selected as the target point cloud. For all neighborhood points within the neighborhood with the target point cloud as the center and the radius as the initial clustering radius of the target point cloud, the cosine similarity value of the three-dimensional velocity sequence of the target point cloud and the three-dimensional velocity sequence of each neighborhood point is calculated as the velocity consistency between the target point cloud and each neighborhood point, and the Euclidean distance between the target point cloud and each neighborhood point is calculated.

[0066] The cosine similarity and Euclidean distance are further combined to comprehensively evaluate whether the target point cloud and each neighboring point belong to the same object, that is, the consistency of speed and spatial distance are taken into account, and then the similarity value between the target point cloud and the neighboring points is calculated, that is, the relationship is satisfied:

[0067]

[0068] Where, For the The target point cloud and the The similarity value of the neighborhood point cloud, For the The target point cloud and the The velocity consistency between neighboring point clouds, For the The target point cloud and the The Euclidean distance between neighboring point clouds, It is an exponential function with the natural constant e as its base.

[0069] It should be noted that the calculation method of cosine similarity and Euclidean distance is an existing technology and will not be described in detail here.

[0070] Furthermore, the initial clustering radius of the target point cloud is adjusted according to the similarity value between the target point cloud and the neighboring points to obtain the optimal clustering radius, which satisfies the relationship:

[0071]

[0072] Where, For the The optimal clustering radius of the target point cloud, For the The initial clustering radius of the target point cloud, For the The target point cloud and the The similarity value of the neighborhood point cloud, is the total number of neighborhood points of the target point cloud, It is an exponential function with the natural constant e as its base.

[0073] When the mean similarity between the target point cloud and all the neighboring points in its neighborhood is negative, it means that the target point cloud and the neighboring points are quite different. If the initial clustering radius is not adjusted, these point cloud data with large differences may be mistakenly clustered together, resulting in mis-segmentation. By reducing the initial clustering radius, the clustering range can be reduced, thereby avoiding merging dissimilar point clouds into the same cluster and improving clustering accuracy. When the mean similarity between the target point cloud and all the neighboring points in its neighborhood is positive, it means that the target point cloud and the neighboring points are highly similar, and these points are likely to belong to the same cluster. By increasing the initial clustering radius, these similar points can be completely clustered together, avoiding the situation where the cluster is divided into multiple small clusters due to a radius that is too small, thereby enhancing the integrity of the clustering.

[0074] According to the above operation of dynamically adjusting the initial distance class radius, the clustering algorithm can better adapt to the local structural changes of the data, and can find more reasonable cluster boundaries in both high-density and low-density areas.

[0075] Further, according to the calculation steps of the optimal clustering radius of the target point cloud mentioned above, the optimal clustering radius corresponding to all other point clouds can be obtained in the same way.

[0076] In another embodiment, the The target point cloud and the The similarity value of the neighborhood point cloud also satisfies the relationship:

[0077]

[0078] Where, For the The target point cloud and the The similarity value of the neighborhood point cloud, For the The target point cloud and the The velocity consistency between neighboring point clouds, For the The target point cloud and the The Euclidean distance between neighboring point clouds, is an exponential function with the natural constant e as the base, For the The target point cloud and the The shape similarity between neighboring point clouds.

[0079] The above shape similarity is obtained and analyzed by The target point cloud and The shape histogram of the neighborhood point cloud is obtained.

[0080] S4: Clustering the point cloud data of a single frame into different clusters using the optimal clustering radius, where each cluster represents a target. Calculating the distance between the target and the chute, comparing the distance with a preset threshold, and triggering an anti-collision warning if the distance is less than the threshold.

[0081] In one embodiment, it should be noted that the cluster radius baseline value set in S2 is a fixed starting point. It may be well adapted to point cloud data in some cases, but in many cases it may not fully reflect the complexity and dynamic changes of point cloud data. Even after the local adjustment in S3, this adjustment based on the fixed baseline value may still have deviations.

[0082] For example, the point cloud data of all frames have similar characteristics overall. However, the point cloud data may have significant differences between different frames, such as density distribution, noise level, object shape, etc. Therefore, relying solely on the initial benchmark value may not globally optimally adapt to the characteristics of all frames.

[0083] By dynamically adjusting the cluster radius baseline, we can globally correct the locally adjusted results. This global correction ensures that clustering results are consistent and comparable across frames. In other words, local adjustments focus on the "characteristics within the point cloud," while global adjustments focus on the "relationships between point clouds."

[0084] Specifically, the optimal clustering radius is used to segment the point cloud data of a single frame through the DBSCAN clustering algorithm to obtain multiple clusters. Each cluster represents a target (such as obstacles such as the cabin wall). The minimum number of points here is set to 5, which means that a cluster needs to contain at least 5 points.

[0085] The ICP (Iterative Closest Point) algorithm is used to align (register) two consecutive frames of point cloud data in order to compare their clustering results.

[0086] For each cluster in the first frame, the degree of aggregation of the corresponding point cloud data in the second frame is calculated, that is, the information entropy of the clustering labels of the point cloud data corresponding to the point cloud data in each cluster in the first frame in the second frame is calculated. The higher the information entropy, the more dispersed the distribution of these points in the point cloud data of the second frame, that is, the less consistent the clustering results between the two frames; conversely, the lower the information entropy, the more consistent the clustering results between the two frames.

[0087] Furthermore, an objective function is established, the sum of all information entropies is input, and the minimum value of the objective function is solved through the optimization algorithm to obtain the optimal clustering radius benchmark value.

[0088] Furthermore, the optimal cluster radius reference value is applied to the above-mentioned related formulas of S2 and S3 to finally obtain the adjusted optimal cluster radius.

[0089] Among them, the expression of the above objective function is:

[0090]

[0091] Where, is the objective function, The first frame The information entropy of the cluster labels corresponding to the point cloud data in the cluster cluster in the second frame, is the number of clusters in the first frame, is the optimal clustering radius benchmark value.

[0092] After obtaining the adjusted optimal clustering radius, the adjusted clustering radius is used to cluster and segment each frame of point cloud data, identify target objects (ships, cabins, cargo, etc.), further calculate the Euclidean distance between the target and the chute, and compare the Euclidean distance with the pre-set safety distance threshold. If the Euclidean distance is less than the safety distance threshold, the collision avoidance warning is triggered.

[0093] The system includes a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the intelligent anti-collision warning method for a ship loader according to the first aspect of the present invention is implemented.

[0094] The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and therefore will not be described in detail here.

[0095] It should be noted that those skilled in the art may make various modifications and improvements without departing from the scope of the present invention, and these modifications and improvements fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be based on the appended claims.

Claims

1. An intelligent anti-collision warning method for a ship loader, characterized in that: include: Acquire pre-processed continuous multi-frame point cloud data of the ship loader and a certain area around it, wherein the certain area around it includes the chute, the cabin, and the cargo; Calculating the local density of each point cloud data, and setting the initial clustering radius according to the local density; Select any point cloud data as the target point cloud, obtain all neighboring point clouds within the initial clustering radius of the target point cloud, calculate the similarity value between the target point cloud and each of the neighboring point clouds, use the similarity value to adjust the initial clustering radius of the target point cloud, obtain the optimal clustering radius of the target point cloud, and then obtain the optimal clustering radius of all point cloud data; the similarity value is positively correlated with the speed consistency between the target point cloud and its neighboring point clouds, and is negatively correlated with the distance between the target point cloud and the neighboring point clouds; The optimal clustering radius is used to cluster the point cloud data of a single frame into different clusters, each cluster represents a target, and the distance between the target and the chute is calculated. The distance is compared with a preset threshold. If the distance is less than the threshold, an anti-collision warning is triggered.

2. The intelligent anti-collision warning method for ship loader according to claim 1 is characterized in that: The point cloud data includes three-dimensional coordinates and reflection intensity of the point cloud data.

3. The intelligent anti-collision warning method for ship loader according to claim 2 is characterized in that: The process of obtaining the local density includes: Normalize the three-dimensional coordinates and reflection intensity of the point cloud data, and use kernel density estimation to calculate the local density of each point cloud data; The Gaussian kernel function is selected and the bandwidth of the kernel density estimate is calculated.

4. The intelligent anti-collision warning method for ship loader according to claim 3 is characterized in that: The initial cluster radius satisfies the relationship: Where, For the The initial clustering radius of point cloud data, is the cluster radius benchmark value, is the global density of all point cloud data, For the The local density of all point cloud data; the global density of all point cloud data is the average of the local densities of all point cloud data.

5. The intelligent anti-collision warning method for ship loader according to claim 4 is characterized in that: The similarity satisfies the relationship: Where, For the The target point cloud and the The similarity value of the neighborhood point cloud, For the The target point cloud and the The velocity consistency between neighboring point clouds, For the The target point cloud and the The Euclidean distance between neighboring point clouds, is an exponential function with the natural constant e as the base; By calculating the The target point cloud and the The cosine similarity between the velocity vectors of the neighborhood point clouds is obtained.

6. The intelligent anti-collision warning method for ship loader according to claim 5 is characterized in that: The optimal clustering radius satisfies the relationship: Where, For the The optimal clustering radius of the target point cloud, For the The initial clustering radius of the target point cloud, For the The target point cloud and the The similarity value of the neighborhood point cloud, is the total number of neighborhood points of the target point cloud, It is an exponential function with the natural constant e as its base.

7. The intelligent anti-collision warning method for ship loader according to claim 6 is characterized in that: The method of obtaining the optimal clustering radius of all point cloud data further includes: Use the optimal clustering radius to segment the point cloud data of a single frame to obtain multiple clusters; select any two consecutive frames of point cloud data and align them; For each point cloud data in the first frame, calculate the information entropy of the cluster label of the corresponding point cloud data in the second frame; An objective function is established, the sum of all information entropies is input, and the optimal clustering radius benchmark value is output; based on the optimal clustering radius benchmark value, the adjusted optimal clustering radius is obtained.

8. The intelligent anti-collision warning method for ship loader according to claim 7 is characterized in that: The expression of the objective function is: Where, is the objective function, The first frame The information entropy of the cluster labels corresponding to the point cloud data in the cluster cluster in the second frame, is the number of clusters in the first frame, is the optimal clustering radius benchmark value.

9. The intelligent anti-collision warning method for ship loader according to claim 8, characterized in that: The clustering algorithm is the DBSCAN clustering algorithm.

10. An intelligent anti-collision warning system for ship loader, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent anti-collision warning method for a ship loader according to any one of claims 1 to 9 is implemented.

Images