Intelligent handling system and handling equipment based on optical communication

Abstract

The present invention discloses an intelligent handling system and handling equipment based on optical communication, which relates to the field of intelligent handling technology, and comprises the following steps: a data acquisition module: used for acquiring images of cargo landing areas in each grid area and performing preprocessing; a model construction module: used for inputting the training set into a target detection model for training to obtain a light landing point detection model; a deviation calculation module: used for calculating the optical signal landing point position deviation coefficient of cargo in each grid area; a path evaluation module: used for calculating a handling path evaluation value, and sorting in ascending order to obtain a handling path sorting table; a path selection module: used for obtaining a candidate handling path set based on the handling path sorting table and a preset maximum deviation value threshold, and calculating the actual handling path length of each candidate handling path for handling path selection, thereby solving the problems of inaccurate reflector positioning, optical signal deviation, and unreasonable handling path selection.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transport technology, and more specifically, to an intelligent transport system and transport equipment based on optical communication. Background Art

[0002] In current intelligent handling systems, optical communications are usually used for the rapid transmission of information such as control instructions, path planning, and real-time feedback. However, there may be various obstacles, narrow areas, or blind spots in the handling environment. The reflector system can achieve multi-point coverage and directional reflection of optical signals, thereby optimizing the signal transmission path and making optical communications more adaptable in complex handling environments. However, when the reflectors are arranged in key areas, the reflectors must be precisely positioned and aligned with the target area to ensure that the optical signal can be accurately reflected to the specified position. A slight deviation may cause the signal to be inaccurate or even fail to reflect to the target area. In addition, when detecting the landing point of the optical signal, a large deviation value will appear, thereby affecting the selection of the handling path. In view of the above problems, the present invention proposes a solution. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent handling system and handling equipment based on optical communication, which solves the problems of inaccurate positioning of the reflector, optical signal deviation, and unreasonable selection of the handling path.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] An intelligent handling system based on optical communication comprises a data acquisition module, a model construction module, a deviation calculation module, a path evaluation module and a path selection module, wherein the modules are connected: a data acquisition module is used to divide a cargo handling area into a plurality of small grid areas, and obtain an image of a cargo landing area of ​​each grid area and perform preprocessing; a model construction module is used to mark the landing position of a light signal in the image of the cargo landing area to obtain a data set, and input the training set into a target detection model for training to obtain a light landing point detection model; a deviation calculation module is used to input the image to be identified of the cargo landing area of ​​each grid area into the light landing point detection model to obtain the light signal landing point position, and calculate the light signal landing point position deviation coefficient of the cargo in each grid area; a path evaluation module is used to calculate a handling path evaluation value based on the light signal landing point position deviation coefficient of the cargo in each grid area, and sort them in ascending order to obtain a handling path sorting table; a path selection module is used to obtain a candidate handling path set based on the handling path sorting table and a preset maximum deviation value threshold, and calculate the actual handling path length of each candidate handling path for handling path selection.

[0006] In a preferred embodiment, the landing position of the light signal is annotated in the image of the cargo landing area to obtain a data set, and the data set is input into the target detection model for training to obtain a light landing point detection model, specifically: the light signal landing point area in the image of the cargo landing area is annotated using a preset annotation shape to obtain a light landing point annotation box; a "light landing point" label is added to each light landing point annotation box, and the annotation information of the image is saved in a preset training set format to obtain a data set;

[0007] The data set is divided into a training set and a test set according to a preset ratio, the training set is input into the target detection model for model training, and the binary cross entropy loss function is selected as the loss function of the model to obtain the trained target detection model; the test set is used to verify the trained target detection model to obtain the light landing point detection model.

[0008] In a preferred embodiment, the image to be identified of the cargo landing area of ​​each grid area is input into the light landing point detection model to obtain the light signal landing point position, specifically: the light landing point detection model selects classic target detection algorithms including YOLO series target detection methods and Faster-RCNN detection algorithms; the classic target detection algorithms are evaluated based on model evaluation indicators, and the model with the best model evaluation effect is selected as the light landing point detection model, and the model evaluation indicators include accuracy, IoU value, false alarm rate and false negative rate;

[0009] In a preferred embodiment, a transport path evaluation value is calculated based on the optical signal landing point position deviation coefficient of the goods in each grid area, and the transport path sorting table is obtained by sorting them in ascending order, specifically: adding the signal landing point position deviation coefficients of the goods in n grid areas in each transport path to obtain a total transport path deviation evaluation value; performing weighted averaging on the total transport path deviation evaluation values ​​to obtain a transport path evaluation value; and sorting the transport path evaluation values ​​in ascending order to obtain a transport path sorting table.

[0010] In a preferred embodiment, based on the transport path sorting table and a preset maximum deviation value threshold, a candidate transport path set is obtained, and the actual transport path length of each candidate transport path is calculated to perform transport path selection, specifically: transport paths in the transport path sorting table whose transport evaluation values ​​are less than a preset minimum transport path evaluation value threshold are taken as candidate transport path sets; based on the candidate transport path set, the actual transport path length of each candidate transport path is calculated; candidate transport paths whose actual transport path lengths are greater than a preset maximum allowable path length threshold are eliminated to obtain a screened valid transport path set; and transport path selection is performed based on the screened valid transport path set.

[0011] In a preferred embodiment, based on the screened effective transport path set, transport path selection is performed, specifically: a comprehensive evaluation index of each effective transport path in the screened effective transport path set is calculated based on a preset transport path comprehensive scoring function; based on the comprehensive evaluation index of each effective transport path, the effective transport paths are sorted, and the effective transport path with the highest comprehensive evaluation index is selected as the optimal transport path.

[0012] The technical effects and advantages of the intelligent handling system and handling equipment based on optical communication of the present invention are as follows:

[0013] 1. The present invention realizes high-precision path planning and selection through the orderly coordination of multiple modules, and has the following advantages: First, the data acquisition module grids the handling area to ensure the fine acquisition and preprocessing of each grid image, thereby improving the accuracy of the data; secondly, the model construction module trains the light landing point detection model through the target detection model, which can accurately identify the landing point of the light signal, making the subsequent deviation calculation more reliable. The deviation calculation module uses the light signal landing point position data to calculate the deviation coefficient of each grid area, providing detailed basic data for path evaluation. The path evaluation module comprehensively sorts the paths according to the deviation coefficient to ensure that high-precision paths are prioritized. Finally, the path selection module screens the candidate path set based on the path sorting table, and comprehensively evaluates it in combination with the actual handling length to ensure that the optimal path is selected under the premise of safety and reliability. Therefore, the system is significantly optimized in terms of accuracy, stability and efficiency, and is particularly suitable for use in handling environments with high precision requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The figure is a schematic diagram of the structure of an intelligent handling system based on optical communication of the present invention. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] Embodiment 1, Figure 1 The present invention provides an intelligent handling system based on optical communication, including a data acquisition module, a model building module, a deviation calculation module, a path evaluation module and a path selection module, and there are connections between the modules:

[0017] Data acquisition module: used to divide the cargo handling area into several small grid areas, and obtain the image of the cargo placement area in each grid area and perform preprocessing;

[0018] In this example, the cargo handling area is divided into several small grid areas, specifically:

[0019] Identify the boundaries of the entire cargo handling area and set the length and width of the area;

[0020] Set the size of each grid according to the size of the cargo and the accuracy requirements of the light signal landing point distribution;

[0021] Based on the size of each grid, the transport area is evenly divided into a number of small grid areas.

[0022] It should be noted that dividing the cargo handling area into small grid areas has many advantages. First, it significantly improves the accuracy of light signal landing point detection, because the environmental characteristics within each grid are relatively consistent, which is conducive to the system accurately capturing the slight deviation of the light signal landing point. Secondly, grid division facilitates partition management and analysis, so that different areas can be handled independently in complex environments, which is convenient for eliminating local interference factors. Thirdly, small grid division provides a more flexible basis for path planning. By analyzing the deviations of each grid, it can optimize path selection and improve the overall efficiency and stability of the handling system.

[0023] In this example, the image of the cargo placement area in each grid area is obtained and preprocessed, specifically:

[0024] Arrange the camera equipment above or around the handling area to ensure that each grid area is within the camera's field of view;

[0025] Capture images of each grid area to ensure that each image contains the complete cargo placement area of ​​the area;

[0026] The collected images are resized uniformly to make them of the same size to meet the input requirements of the target detection model;

[0027] The image is denoised by using a filter algorithm (such as Gaussian filtering) to reduce the noise in the image, improve the accuracy of model recognition, and adjust the contrast and brightness of the image according to the lighting conditions, so that the light signal landing point in the image is clearer.

[0028] It should be noted that acquiring and preprocessing the cargo landing area images of each grid area has the advantages of accuracy and reliability. First, by acquiring images grid by grid, the system can more accurately locate the landing point of the light signal and improve the detection accuracy. Secondly, preprocessing steps such as image enhancement and denoising can help eliminate noise caused by light changes or background interference, ensuring image quality and the accuracy of subsequent analysis. In addition, this regional image acquisition method makes the system more targeted when monitoring and optimizing each grid, making it easier to identify abnormal landing points and optimize the handling path, ultimately improving the stability and efficiency of the intelligent handling system.

[0029] Model building module: used for marking the landing position of the light signal in the image of the cargo landing area to obtain a data set, and inputting the training set into the target detection model for training to obtain a light landing point detection model;

[0030] In this example, the landing position of the light signal is marked in the image of the cargo landing area to obtain a data set, and the data set is input into the target detection model for training to obtain a light landing point detection model, specifically:

[0031] Annotate the light signal landing point area in the image of the cargo landing area using a preset annotation shape to obtain a light landing point annotation frame;

[0032] Add a "light point" label to each light point annotation box, and save the image annotation information in a preset training set format to obtain a data set;

[0033] The data set is divided into a training set and a test set according to a preset ratio, the training set is input into the target detection model for model training, and a binary cross entropy loss function is selected as the loss function of the model to obtain a trained target detection model;

[0034] The trained target detection model is verified using a test set to obtain a light landing point detection model.

[0035] It should be noted that the method of marking the light signal landing points in the cargo landing area image and training the target detection model has the advantages of high efficiency, accuracy and strong applicability. By marking the light signal landing points with preset shapes and adding labels, a standardized data set can be systematically formed to ensure the consistency of training samples. At the same time, the binary cross entropy loss function is used for training to improve the model's detection accuracy of light landing points. The verification phase of the test set provides effective performance feedback for the model, making the final model more robust. The overall method ensures the model's precise positioning capability, which is conducive to improving the reliability and path optimization effect of the optical communication cargo handling system.

[0036] Deviation calculation module: used to input the image to be identified of the cargo landing area of ​​each grid area into the light landing point detection model, obtain the light signal landing point position, and calculate the light signal landing point position deviation coefficient of the cargo in each grid area;

[0037] In this example, the light falling point detection model selects classic target detection algorithms including YOLO series target detection methods and Faster-RCNN detection algorithms;

[0038] The classic target detection algorithm is evaluated based on model evaluation indicators, and the model with the best model evaluation effect is selected as the light point detection model. The model evaluation indicators include accuracy, IoU value, false alarm rate and missed alarm rate.

[0039] In this example, the classic target detection algorithms include the YOLO series target detection method and the Faster R-CNN detection algorithm, specifically:

[0040] YOLO is a single-stage detection algorithm that takes the entire image as input and directly predicts the category and location of multiple objects. YOLO divides the image into grids, each of which is responsible for predicting one or more bounding boxes and category probabilities. YOLO v3, v4 and the latest YOLO v5, YOLOv7 and other series versions are continuously optimized in detection accuracy and speed, and the detection performance of small targets is improved by improving the network structure (such as CSPNet, PANet, etc.) and using multi-scale feature fusion. The advantage of the YOLO model is its fast detection speed and good real-time performance, which is suitable for scenarios that require fast response, such as light signal landing point detection in automated handling path optimization;

[0041] Faster R-CNN is a two-stage detection method that evolved from the R-CNN series. Faster R-CNN first extracts features through a convolutional neural network (CNN), and then generates candidate regions (Region Proposals) to locate possible object regions. The key innovation of this algorithm is the introduction of RPN (Region Proposal Network), which greatly speeds up the speed of region proposals. Subsequently, Faster R-CNN further classifies these regions and regresses the precise bounding box positions. This algorithm performs well in detection accuracy, but compared to YOLO, it has a slightly larger amount of computation and is suitable for scenarios with high detection accuracy requirements.

[0042] In this example, the optical signal landing point position deviation coefficient of the goods in each grid area is calculated. The specific formula is as follows:

[0043]

[0044] in, is the light signal landing point position deviation coefficient of the cargo in each grid area, , and are the three-dimensional coordinates of the actual light signal landing point, , and They are the ideal landing positions of the signal under interference-free conditions, is the optical signal quality correction factor.

[0045] It should be noted that the advantage of calculating the deviation coefficient of the optical signal landing point position of the goods in each grid area is that it can accurately evaluate and optimize the coverage effect of the optical signal in each area. By quantifying the optical signal deviation in each grid area, the system can monitor and adjust the propagation accuracy of the optical signal in real time to ensure the quality of optical communication on the cargo handling path. This method helps to reduce signal loss and interference and improve the reliability and stability of the optical signal. At the same time, the calculation of the deviation coefficient can help identify areas with unstable signals or incomplete coverage, making it easier to adjust the position of the reflector or add light source equipment, thereby optimizing the performance of the entire system. Ultimately, this precise deviation analysis provides a basis for path selection, helps to avoid signal attenuation and misdirection during handling, and improves handling efficiency and safety.

[0046] Path evaluation module: used to calculate the transport path evaluation value based on the optical signal landing point position deviation coefficient of the goods in each grid area, and sort them in ascending order to obtain a transport path sorting table;

[0047] In this example, the transport path evaluation value is calculated based on the optical signal landing point position deviation coefficient of the goods in each grid area, and the transport path sorting table is obtained by sorting in ascending order, specifically:

[0048] Add the signal landing point position deviation coefficients of the goods in n grid areas in each transport path to obtain the total deviation evaluation value of the transport path;

[0049] Taking a weighted average of the total deviation evaluation values ​​of the transport path to obtain a transport path evaluation value;

[0050] The transport path evaluation values ​​are sorted in ascending order to obtain a transport path sorting table.

[0051] In this example, the signal drop point position deviation coefficients of the goods in n grid areas in each transport path are added together to obtain the total deviation evaluation value of the transport path, which is specifically:

[0052]

[0053] in, is the total deviation evaluation value of the transport path, is the signal landing point position deviation coefficient of the cargo in the i-th grid area, indicating the position error of the signal in this area deviating from the target landing point, and n is the total number of grid areas passed on the transport path.

[0054] In this example, the total deviation evaluation value of the transport path is weighted averaged to obtain the transport path evaluation value. The specific formula is as follows:

[0055]

[0056] in, is the transport path evaluation value, is the path length of each transport path, is the total number of grid areas passed in the path, is the optical signal deviation coefficient of the i-th grid area.

[0057] It should be noted that the scheme of calculating the transport path evaluation value based on the optical signal landing point position deviation coefficient of the goods in each grid area and sorting them in ascending order has significant advantages. First, by adding the signal landing point position deviation coefficients of each grid area in each transport path, the signal quality of each area on the path can be comprehensively measured to ensure the overall communication stability of the path. Secondly, the importance of different areas in path evaluation is further considered through the weighted average method to ensure that the signal quality of key areas can be given priority attention, thereby improving the accuracy of the evaluation results. Finally, the transport path sorting table is obtained by sorting in ascending order, which can provide a quantitative decision-making basis for the system, so that the system can give priority to paths with better signal quality, optimize the handling efficiency, and avoid communication problems or delays caused by selecting paths with poor signal quality. This method effectively improves the intelligence and scientificity of path selection and ensures the efficient and safe operation of the system.

[0058] Path selection module: used to obtain a set of candidate transport paths based on the transport path sorting table and a preset maximum deviation value threshold, and calculate the actual transport path length of each candidate transport path to select the transport path.

[0059] In this example, based on the transport path sorting table and a preset maximum deviation value threshold, a candidate transport path set is obtained, and the actual transport path length of each candidate transport path is calculated to select a transport path, specifically:

[0060] The transport paths in the transport path ranking table whose transport evaluation values ​​are less than a preset minimum transport path evaluation value threshold are selected as candidate transport path sets;

[0061] Based on the candidate transport path set, calculating an actual transport path length of each candidate transport path;

[0062] Eliminate the candidate transport paths whose actual transport path length is greater than a preset maximum allowable path length threshold, to obtain a screened valid transport path set;

[0063] Based on the filtered effective transport path set, transport path selection is performed.

[0064] In this example, based on the candidate transport path set, the actual transport path length of each candidate transport path is calculated, specifically:

[0065] For each candidate transport path, determine the center point of each grid area on the path;

[0066] The Manhattan distance calculation formula is used to calculate the length between the center points of adjacent grid areas;

[0067] The lengths between the center points of adjacent grid areas are added to obtain the actual transport path length of each candidate transport path.

[0068] In this example, the Manhattan distance calculation formula is specifically:

[0069]

[0070] in, is the length between the center points of adjacent grid areas, and are the horizontal and vertical coordinates of the center point of the i-th grid area, and The horizontal and vertical coordinates of the center points of adjacent grid areas respectively.

[0071] It should be noted that using the Manhattan distance calculation formula to calculate the length between the center points of adjacent grid areas has several practical advantages. First, the Manhattan distance is suitable for grid environments, where movement is usually limited to the horizontal and vertical directions. This calculation method meets the path selection requirements in actual handling scenarios, especially when path planning is subject to physical restrictions (such as obstacles, walls, etc.). Manhattan distance provides a concise and efficient calculation method because it directly calculates the path along the vertical and horizontal lines of the grid without considering the diagonal direction. Therefore, Manhattan distance can avoid complex calculations while reducing errors in the calculation process.

[0072] Secondly, Manhattan distance is fast to calculate, especially in large-scale grid environments, avoiding the complexity of square root calculations like Euclidean distance. This makes it possible to use Manhattan distance in real-time path planning to improve response speed and adapt to dynamically changing cargo handling scenarios.

[0073] In addition, the calculation method of Manhattan distance can better simulate obstacles and path restrictions in actual situations. In some warehouse environments or handling scenarios, the robot's moving path is often restricted by fixed channels or walls, and Manhattan distance can provide reasonable path evaluation in these scenarios. Therefore, in the actual intelligent handling system, the use of Manhattan distance to calculate the length between the center points of adjacent grid areas not only improves the calculation efficiency, but also better meets the requirements of the physical environment, making the path selection more in line with the actual operation needs.

[0074] In this example, based on the filtered effective transport path set, transport path selection is performed, specifically:

[0075] Calculating a comprehensive evaluation index of each effective transport path in the screened effective transport path set based on a preset transport path comprehensive scoring function;

[0076] Based on the comprehensive evaluation index of each effective transport path, the effective transport paths are sorted, and the effective transport path with the highest comprehensive evaluation index is selected as the optimal transport path.

[0077] In this example, the comprehensive evaluation index of each effective transport path in the screened effective transport path set is calculated based on a preset comprehensive scoring function of the transport path. The specific formula is as follows:

[0078]

[0079] in, is the comprehensive evaluation index for each effective transport path, is the preset maximum allowed path length threshold, is the length of each effective transport path, is the optical signal landing point position deviation coefficient threshold, is the optical signal landing position deviation coefficient for each effective transport path, is the environmental impact coefficient of each effective transport path, , and is the weight coefficient.

[0080] It should be noted that the selection of transport paths based on the screened effective transport path set, the use of a preset comprehensive scoring function for transport paths, and the calculation of the comprehensive evaluation index of each effective transport path have the following significant advantages. First, by introducing a comprehensive scoring function, multiple factors can be comprehensively considered, such as the optical signal deviation of the path, the path length, environmental impact, etc., to provide a comprehensive evaluation standard. This method can comprehensively reflect the overall performance of the path, avoid the one-sidedness of a single factor, and make the path selection more scientific and reasonable.

[0081] Secondly, the comprehensive scoring function can flexibly adjust the priority of different factors by weighting various indicators, so as to make personalized adjustments according to actual needs. For example, in some scenarios, the optical signal stability of the path may be more important than the path length. By adjusting the weight of the scoring function, the selection result can be more in line with the actual transportation needs.

[0082] Third, by sorting the comprehensive evaluation index of each effective transport path, the optimal path can be determined intuitively and human intervention and errors can be reduced. The sorting mechanism ensures the objectivity and automation of path selection, enabling the system to make quick decisions based on real-time data, improving the efficiency and response speed of the intelligent transport system.

[0083] In summary, this path selection method based on the comprehensive scoring function not only improves the intelligence and automation level of path selection, but also effectively balances multiple factors, optimizes transportation efficiency, and improves the overall performance of the system.

[0084] The present invention also includes an electronic device, the electronic device comprising:

[0085] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute an intelligent handling system based on optical communication.

[0086] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0087] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0088] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0089] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0090] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0091] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent handling system based on optical communication, characterized in that: It includes data acquisition module, model building module, deviation calculation module, path evaluation module and path selection module. There are connections between modules: Data acquisition module: used to divide the cargo handling area into several small grid areas, and obtain the image of the cargo placement area in each grid area and perform preprocessing; Model building module: used for marking the landing position of the light signal in the image of the cargo landing area to obtain a data set, and inputting the data set into the target detection model for training to obtain a light landing point detection model; Deviation calculation module: used to input the image to be identified of the cargo landing area of ​​each grid area into the light landing point detection model, obtain the light signal landing point position, and calculate the light signal landing point position deviation coefficient of the cargo in each grid area; Calculate the optical signal landing point position deviation coefficient of the goods in each grid area. The specific formula is as follows: in, is the light signal landing point position deviation coefficient of the cargo in each grid area, , and are the three-dimensional coordinates of the actual light signal landing point, , and They are the ideal landing positions of the signal under interference-free conditions, is the optical signal quality correction factor; Path evaluation module: used to calculate the transport path evaluation value based on the optical signal landing point position deviation coefficient of the goods in each grid area, and sort them in ascending order to obtain a transport path sorting table; Add the signal landing point position deviation coefficients of the goods in n grid areas in each transport path to obtain the total deviation evaluation value of the transport path; Taking a weighted average of the total deviation evaluation values ​​of the transport path to obtain a transport path evaluation value; Sorting the transport path evaluation values ​​in ascending order to obtain a transport path sorting table; Path selection module: used to obtain a set of candidate transport paths based on the transport path sorting table and a preset maximum deviation value threshold, and calculate the actual transport path length of each candidate transport path to select the transport path, specifically: The transport paths in the transport path ranking table whose transport evaluation values ​​are less than a preset minimum transport path evaluation value threshold are selected as candidate transport path sets; Based on the candidate transport path set, calculating an actual transport path length of each candidate transport path; Eliminate the candidate transport paths whose actual transport path length is greater than a preset maximum allowable path length threshold, to obtain a screened valid transport path set; Based on the screened effective transport path set, transport path selection is performed, specifically: Calculating a comprehensive evaluation index of each effective transport path in the screened effective transport path set based on a preset transport path comprehensive scoring function; The comprehensive evaluation index of each effective transport path in the screened effective transport path set is calculated based on the preset transport path comprehensive scoring function. The specific formula is as follows: in, is the comprehensive evaluation index for each effective transport path, is the preset maximum allowed path length threshold, is the length of each effective transport path, is the optical signal landing point position deviation coefficient threshold, is the optical signal landing position deviation coefficient for each effective transport path, is the environmental impact coefficient of each effective transport path, , and is the weight coefficient; Based on the comprehensive evaluation index of each effective transport path, the effective transport paths are sorted, and the effective transport path with the highest comprehensive evaluation index is selected as the optimal transport path.

2. The intelligent handling system based on optical communication according to claim 1, characterized in that: The landing point positions of the light signals are marked in the image of the cargo landing area to obtain a data set, and the data set is input into the target detection model for training to obtain a light landing point detection model, specifically: Annotate the light signal landing point area in the image of the cargo landing area using a preset annotation shape to obtain a light landing point annotation frame; Add a light point label to each light point annotation box, and save the image annotation information in a preset training set format to obtain a data set; The data set is divided into a training set and a test set according to a preset ratio, the training set is input into the target detection model for model training, and a binary cross entropy loss function is selected as the loss function of the model to obtain a trained target detection model; The trained target detection model is verified using a test set to obtain a light landing point detection model.

3. The intelligent handling system based on optical communication according to claim 2, characterized in that: The image to be identified of the cargo landing area in each grid area is input into the light landing point detection model to obtain the light signal landing point position, specifically: The light falling point detection model selects classic target detection algorithms including YOLO series target detection methods and FasterR-CNN detection algorithm; The classic target detection algorithm is evaluated based on model evaluation indicators, and the model with the best model evaluation effect is selected as the light point detection model. The model evaluation indicators include accuracy, IoU value, false alarm rate and missed alarm rate.

4. The intelligent handling system based on optical communication according to claim 3, characterized in that: The total deviation evaluation value of the transport path is weighted averaged to obtain the transport path evaluation value. The specific formula is as follows: in, is the transport path evaluation value, is the path length of each transport path, is the total number of grid areas passed in the path, is the optical signal deviation coefficient of the i-th grid area.

5. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute an intelligent handling system based on optical communication as described in any one of claims 1 to 4.

Images