Cargo height and width detection system and related method and device

Abstract

The invention discloses a cargo height and width detection system and a related method and device. The method or system comprises the steps that the distance between the AGV forklift and a front operation area is monitored in real time, and image collection is triggered when the distance reaches a preset activation threshold value; carrying out edge feature extraction and gray level comparison analysis on the obtained two-dimensional digital image, and identifying a cargo shadow area; calculating key coordinate points of the shadow area based on a perspective projection principle, and determining the height and width of the cargo through a geometric mapping relation; when the detected size exceeds a safety threshold value, a grading early warning signal is generated, and sudden stop is controlled. According to the method, the energy consumption of the sensor is reduced through a dynamic trigger mechanism, the problem of misjudgment caused by environmental interference is solved by combining gray scale contrastive analysis and a perspective projection algorithm, and the balance of safety and efficiency is realized by utilizing graded early warning. Compared with the prior art, the method has the advantages that the detection precision and the system reliability are remarkably improved, collision accidents in the cargo carrying process can be effectively prevented, and data support is provided for storage space optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent warehousing logistics, and in particular to a cargo overheight and overwidth detection system and related methods and devices. BACKGROUND

[0002] In the field of modern logistics and warehousing, automated guided vehicle (AGV) forklifts, as the core equipment for intelligent material handling, have significantly improved operational efficiency and effectively reduced labor costs. With the rapid development of e-commerce and manufacturing, the specifications of goods in the logistics and warehousing scenario are becoming increasingly diverse, and the demand for handling overheight and overwidth goods is becoming increasingly prominent. This trend has placed higher demands on the cargo identification and handling capabilities of AGV forklifts, making the AGV forklift cargo overheight and overwidth detection system a key technical support for ensuring operational safety and improving the utilization rate of warehouse space. Accurate cargo size detection not only effectively prevents collisions from occurring, but also provides data support for the optimal layout of warehouse space, thereby promoting the overall efficiency of the intelligent warehousing system.

[0003] However, the cargo detection system of AGV forklifts in the prior art still has many technical bottlenecks that need to be optimized. The commonly used multi-sensor fusion scheme, which involves distributed deployment of multiple sensors or cameras on the vehicle body for joint detection, can achieve basic size detection functions, but the multi-sensor fusion detection system of existing AGV forklifts uses a non-discriminatory continuous working mode, resulting in significant resource waste and a high susceptibility to environmental interference, which affects system reliability. These technical limitations to some extent restrict the popularization and application of AGV forklifts in complex warehouse environments, and urgent technical innovation is needed to address these issues. SUMMARY

[0004] The present application aims to solve the technical bottlenecks of severe system resource waste, high environmental interference sensitivity, and insufficient detection reliability in existing AGV forklift cargo overheight and overwidth detection technology. By innovatively proposing an intelligent detection mechanism based on job spacing triggering, the detection system is activated on demand and operates accurately, significantly reducing system power consumption while improving detection accuracy, ultimately building an intelligent detection system with environmental adaptability, providing reliable protection for safe operation in complex warehouse environments.

[0005] To achieve the above-mentioned purpose, the first aspect of the present application provides an AGV forklift cargo overheight and overwidth detection system, comprising: a distance measurement and condition judgment module for real-time monitoring of the distance between the AGV forklift and the front work area and determining whether the distance reaches a preset activation threshold; an image acquisition module connected to the distance measurement and condition judgment module, for acquiring a two-dimensional digital image of the target area when the distance reaches the preset activation threshold; The shadow analysis module is connected with the image acquisition module, and is configured to extract edge features of the two-dimensional digital image, and analyze whether a shadow area of the goods is formed in the front operation area through comparison of gray values. The coordinate positioning module is connected with the distance measuring and condition judging module and the shadow analysis module, and is configured to determine the coordinates of the highest point, the lowest point, the leftmost point and the rightmost point of the shadow area based on the principle of perspective projection and the interval when the shadow area of the goods exists. The size calculation module is connected with the coordinate positioning module, and is configured to calculate the height size of the goods according to the coordinates of the highest point and the lowest point, and calculate the width size of the goods according to the horizontal interval of the coordinates of the leftmost point and the rightmost point. The early warning execution module is connected with the size calculation module, and is configured to generate a graded early warning signal and control the AGV forklift to stop urgently when the height size exceeds the first safety threshold and / or the width size exceeds the second safety threshold.

[0006] Further, the shadow analysis module comprises: The pallet recognition unit is configured to recognize whether a target pallet exists in the target area based on the contour features of the preset pallet. The shadow verification unit is connected with the pallet recognition unit, and is configured to analyze whether a shadow area of the goods is formed in a predetermined area at the top of the target pallet through comparison of gray values when the target pallet is recognized.

[0007] Further, the pallet recognition unit comprises: The preprocessing subunit is configured to perform preprocessing on the two-dimensional digital image. The feature comparison subunit is connected with the preprocessing subunit, and is configured to compare the preprocessed data with a three-dimensional contour feature database. The protrusion extraction subunit is configured to extract protruding structures through an edge detection algorithm. The size screening subunit is connected with the protrusion extraction subunit, and is configured to screen protruding structures that meet the preset support size specification. The topological verification subunit is connected with the size screening subunit, and is configured to verify whether the relative positions of the protruding structures meet the spatial geometric constraint conditions of the tripod pallet. The output subunit is configured to confirm that the target pallet exists when the size screening and / or topological verification are met.

[0008] Further, the shadow analysis module comprises: The gray scale analysis unit is configured to perform gray scale analysis on a preset detection area in the two-dimensional digital image, and output a shadow area to be verified when the average gray scale is lower than a preset percentage of the background area. The contour extraction unit is connected with the gray scale analysis unit, and is configured to perform edge detection and contour extraction on the shadow area to be verified, and calculate a contour integrity index P ∈ [0, 1]. a gradient calculation unit configured to extract a gradient feature of a pixel at an edge of the shadow region to be verified and calculate a gradient amplitude standard deviation G; a confidence score unit connected to the contour extraction unit and the gradient calculation unit, configured to calculate a confidence score Cscore based on the contour integrity index P and the gradient amplitude standard deviation G, according to the following formula: wherein a and β are preset weight coefficients, and a + β = 1, G max is a preset maximum gradient amplitude standard deviation; a determination unit connected to the confidence score unit, configured to determine that the shadow region of the cargo when C score ≥ a preset score threshold; a thermal imaging verification unit connected to the determination unit, configured to start an infrared thermal imaging module to perform secondary verification when C score < a preset score threshold.

[0009] Further, the shadow analysis module further comprises: an illumination sensing unit configured to measure an ambient light intensity Lux value in a front work area; a dynamic threshold unit connected to the illumination sensing unit, configured to dynamically adjust a gray scale threshold interval based on the Lux value, according to the following formula: wherein β is an illumination adaptation coefficient, Lux base is a calibrated ambient light reference value; a shadow detection unit connected to the dynamic threshold unit, configured to perform shadow region detection based on the gray scale value after the threshold interval is adjusted.

[0010] Further, the system further comprises: a tilt angle sensing module configured to measure a real-time tilt angle θ of the AGV forklift; a height correction module connected to the size calculation module and the tilt angle sensing module, respectively, configured to correct the cargo height size based on the real-time tilt angle θ, according to the following correction formula: wherein, H raW is an original height value, Δh cal is a cargo fork deformation compensation amount.

[0011] Further, the early warning execution module further comprises: a cooperative request unit configured to send a detection request containing shadow region coordinates to i AGV forklifts in the surrounding when the cargo width is out of limit; a data fusion unit connected to the cooperative request unit, configured to receive multi-view image data and confidence scores C score,i of the shadow region of the cargo returned by each AGV forklift in the surrounding; The three-dimensional modeling unit is connected with the data fusion unit and is configured to generate a three-dimensional point cloud model through feature point matching. The weighted calculation unit is configured to perform weighted average calculation on the comprehensive cargo width value according to the confidence score: W i The AGV forklift original width measurement value of each periphery.

[0012] In a second aspect, the application provides an AGV forklift cargo overheight and overwidth detection method, comprising: S100, real-time monitoring the distance between the AGV forklift and the front work area; S200, when the distance reaches a preset activation threshold, acquiring a two-dimensional digital image of the target area; S300, performing edge feature extraction on the two-dimensional digital image, and analyzing whether a cargo shadow area is formed in the front work area through gray value comparison; S400, if the cargo shadow area exists, determining the highest point coordinates, the lowest point coordinates, the leftmost point coordinates and the rightmost point coordinates of the cargo shadow area according to the principle of perspective projection; S500, calculating the cargo height dimension based on the highest point coordinates and the lowest point coordinates, and calculating the cargo width dimension based on the horizontal distance between the leftmost point coordinates and the rightmost point coordinates; S600, when the cargo height dimension exceeds a first safety threshold and / or the cargo width dimension exceeds a second safety threshold, generating a hierarchical early warning signal, and controlling the AGV forklift to perform an emergency stop operation.

[0013] Further, the gray value comparison analysis of whether a cargo shadow area is formed in the front work area specifically comprises: C310, identifying whether a target pallet exists in the target area based on a preset pallet contour feature; C320, when the target pallet is identified, analyzing whether a cargo shadow area is formed in a predetermined area on the upper end of the target pallet through gray value comparison.

[0014] In a third aspect, the application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the steps of the AGV forklift cargo overheight and overwidth detection method.

[0015] In a fourth aspect, the application provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to realize the steps of the AGV forklift cargo overheight and overwidth detection method.

[0016] ​The AGV forklift cargo overheight and overwidth detection system and method provided by the application has at least the following beneficial effects: The application realizes the on-demand activation of the detection system through the work interval induction, breaks through the traditional non-discriminatory continuous monitoring mode, establishes an intelligent triggering mechanism based on the work interval, and makes the detection system only activated under necessary working conditions, so as to fundamentally eliminate the energy loss caused by invalid detection and significantly improve the endurance of the equipment; at the same time, the detection system effectively shields the environmental factors such as light changes and dust interference in the warehouse scene, greatly reduces the probability of false triggering, and improves the reliability of the detection system under different working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A principle diagram of an AGV forklift cargo overheight and overwidth detection system provided by an embodiment of the application is provided; Figure 2 A principle diagram of a shadow analysis module provided by an embodiment of the application is provided; Figure 3 A principle diagram of a tray recognition unit provided by an embodiment of the application is provided; Figure 4 A principle diagram of another shadow analysis module provided by an embodiment of the application is provided; Figure 5 A flowchart of an AGV forklift cargo overheight and overwidth detection method provided by an embodiment of the application is provided; Figure 6 A principle diagram of an electronic device provided by an embodiment of the application is provided.

[0018] REFERENCE NUMERALS Distance measuring and condition judging module-100, image acquisition module-200, shadow analysis module-300, coordinate positioning module-400, size calculation module-500, tray recognition unit-310, shadow verification unit-320, preprocessing subunit-311, feature comparison subunit-312, protrusion extraction subunit-313, size screening subunit-314, topology verification subunit-315, output subunit-316, gray scale analysis unit-330, contour extraction unit-340, gradient calculation unit-350, confidence score unit-360, determination unit-370, thermal imaging verification unit-380, electronic device-600, memory-610, processor-620, computer program-630. DETAILED DESCRIPTION

[0019] With reference to the drawings of the embodiments herein, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0020] Reference herein to“an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase that the phrase in various places in the specification are not necessarily all referring to the same embodiment, or are necessarily mutually exclusive or alternative embodiments. It is explicitly stated that the embodiments described herein can be combined with other embodiments.

[0021] Embodiment one Reference Figure 1 , Figure 1 is a schematic diagram of an AGV forklift cargo overheight and overwidth detection system provided by an embodiment of the present application. The system comprises: a distance measuring and condition judging module 100, configured to monitor the distance between the AGV forklift and the front work area in real time and determine whether the distance reaches a preset activation threshold; an image acquisition module 200, connected to the distance measuring and condition judging module 100, configured to acquire a two-dimensional digital image of the target area when the distance reaches the preset activation threshold; a shadow analysis module 300, connected to the image acquisition module 200, configured to extract edge features from the two-dimensional digital image and analyze whether the front work area forms a cargo shadow area through gray value comparison; a coordinate positioning module 400, connected to the distance measuring and condition judging module 100 and the shadow analysis module 300, respectively, configured to determine the highest point coordinate, the lowest point coordinate, the leftmost point coordinate, and the rightmost point coordinate of the shadow area based on the perspective projection principle and the distance when there is a cargo shadow area; a size calculation module 500, connected to the coordinate positioning module 400, configured to calculate the cargo height size according to the highest point coordinate and the lowest point coordinate, and calculate the cargo width size according to the horizontal distance between the leftmost point coordinate and the rightmost point coordinate; a warning execution module 600, connected to the size calculation module 500, configured to generate a hierarchical warning signal and control the AGV forklift to stop urgently when the height size exceeds a first safety threshold and / or the width size exceeds a second safety threshold.

[0022] In the example embodiment, the "preset activation threshold" refers to a safety distance threshold (usually set in the range of 0.5-3m, and the optimized value in the experiment can be 1.5m) pre-calibrated according to the AGV forklift braking performance parameters and sensor field of view angle characteristics, used to determine the detection system startup timing; the "cargo shadow area" refers to the gray scale transition area formed in the two-dimensional digital image after the cargo blocks the ambient light source, and the boundary features are quantitatively identified by image processing algorithms (typically represented by a gray value that is 20%-50% lower than the background area average, and the embodiment uses 30% as the discrimination threshold); the "perspective projection principle" refers to a geometric transformation method that eliminates the size measurement error caused by the camera installation angle by establishing the mapping relationship between the image coordinate system and the world coordinate system based on the monocular vision imaging model; the "graded warning signal" is a multi-level safety instruction divided according to the cargo over-limit ratio, including three response modes: deceleration warning (over-limit ≤10%), parking warning (10%<over-limit ≤20%), and emergency stop instruction (over-limit >20%); the "highest point coordinate" refers to the maximum coordinate value of the shadow area profile in the vertical direction (world coordinate system Y axis), which corresponds to the top position of the cargo in the three-dimensional space, used to calculate the height dimension of the cargo; the "lowest point coordinate" refers to the minimum coordinate value of the shadow area profile in the vertical direction (world coordinate system Y axis), which corresponds to the contact position of the cargo and the bearing surface, used to calculate the height reference of the cargo; the "leftmost point coordinate" refers to the minimum coordinate value of the shadow area profile in the horizontal direction (world coordinate system X axis), which defines the left boundary of the cargo, used to calculate the width dimension of the cargo; the "rightmost point coordinate" refers to the maximum coordinate value of the shadow area profile in the horizontal direction (world coordinate system X axis). This point defines the right boundary of the cargo, used to calculate the width dimension of the cargo.

[0023] In the example embodiment, the AGV forklift cargo over-height and over-width detection system realizes intelligent safety detection through multi-module cooperation, and its core components include the distance measurement and condition judgment module 100, the image acquisition module 200, the shadow analysis module 300, the coordinate positioning module 400, the size calculation module 500, and the warning execution module 600. The distance measurement and condition judgment module 100 uses a millimeter wave radar or laser radar sensor, for example, to monitor the distance data between the forklift and the front work area in real time at a refresh rate of 50Hz, and triggers the image acquisition process when the monitored distance reaches the preset activation threshold (such as 1.5 meters). The threshold is dynamically calibrated according to the forklift braking performance, for example, by experimentally determining that the braking distance of the forklift in the full load state is 1.2 meters, and setting a safety margin after considering the sensor response delay, finally determining the trigger threshold of 1.5 meters, to ensure that the detection system starts in time when the forklift approaches the potential work area.

[0024] The image acquisition module 200 is configured with an industrial-grade imaging device, such as a global shutter CMOS camera, with a resolution of not less than 2 million pixels, which automatically captures a grayscale image of the work area after receiving a trigger signal. The module integrates an automatic exposure control function, such as adjusting the gain to maintain the image signal-to-noise ratio in a weak light scene, to ensure the recognizability of shadow features. The image transmission uses a gigabit Ethernet interface, with a single-frame data transmission delay controlled within 10 milliseconds, providing a real-time data source for subsequent processing modules.

[0025] The shadow analysis module 300 realizes accurate identification of the cargo shadow area through a two-stage algorithm: first, an improved Canny edge detection algorithm is used to extract image contour features, and then an Otsu adaptive threshold segmentation technique is applied to quantify the grayscale difference. For example, set the adjacent pixel grayscale difference threshold to 30 levels, and when the average grayscale value of the detection area is lower than 30% of the background area, it is determined as an effective shadow area. This design effectively suppresses environmental light interference and maintains detection stability under warehouse scene illumination fluctuations (such as 100-10000 lux range), with a high level of shadow area recognition accuracy.

[0026] The coordinate positioning module 400 realizes space mapping based on a monocular vision geometry model, such as converting the image coordinate system to the world coordinate system through a pre-calibrated perspective transformation matrix. The module receives real-time distance data output by the distance measurement module, and combines with the shadow area contour feature points to calculate the spatial coordinates of the top, bottom, and two sides of the cargo. The coordinate mapping process uses an anti-distortion correction algorithm, such as eliminating lens radial distortion through polynomial fitting, to ensure size measurement accuracy. Experiments show that within an installation angle deviation of ±5°, the system still maintains a coordinate positioning error of less than 5mm.

[0027] The size calculation module 500 performs geometric operations based on coordinate extreme points, such as using the Euclidean distance formula to calculate the vertical distance between the highest point and the lowest point as the cargo height, and the horizontal distance between the leftmost point and the rightmost point as the cargo width. The calculation results are filtered and output, such as applying a sliding average algorithm to suppress transient measurement noise and improve data stability. In a typical application scenario, when the detected cargo height is 2.4 meters (the first safety threshold is set to 2.2 meters), the system automatically triggers an over-limit warning process.

[0028] The Early Warning Execution Module 600 implements a three-level response mechanism: When cargo dimensions exceed a safety threshold, it first sends graded instructions to the AGV controller via the CAN bus. For example, if the dimension exceeds a safety threshold by 10%, a deceleration warning is triggered (reducing the forklift's operating speed from 1.5 m / s to 0.5 m / s); if the dimension exceeds a safety threshold by 10%, a stop warning is initiated (resulting in a forklift stop within 2 seconds); and if the dimension exceeds a safety threshold by more than 20%, an emergency stop is immediately initiated (response time is controlled within 200 milliseconds). This design forms a complete safety closed loop of detection, decision-making, and execution. In experimental testing, it successfully prevented multiple simulated collisions and significantly improved warehouse safety.

[0029] This system achieves millimeter-level detection accuracy in typical warehouse scenarios, significantly reducing false positives through a collaborative mechanism of dynamic triggering and multimodal verification. Experiments have shown that the system maintains high recognition accuracy even in reflective environments or complex backgrounds, making it particularly suitable for automated operations in high-density racking environments.

[0030] This embodiment achieves the following technical effects through the synergy of dynamic triggering and multimodal verification: Adopting the interval threshold trigger mode, it effectively reduces the system energy consumption during non-essential detection periods, forming an intelligent detection mode that works on demand; Combining grayscale gradient analysis and geometric feature verification, it significantly improves the recognition reliability of cargo shadow areas and effectively suppresses ambient light interference. Use perspective projection modeling to eliminate measurement errors caused by camera installation angles and ensure dimensional detection accuracy; Build a hierarchical warning and emergency stop linkage mechanism to form a safety closed loop of detection-decision-execution, and effectively prevent collision accidents in warehousing operations.

[0031] Example 2 Based on the above embodiment 1, this embodiment of the application also provides another AGV forklift cargo overheight and overwidth detection system, see Figure 1 , the system comprises: The distance measurement and condition judgment module 100 is used to monitor the distance between the AGV forklift and the working area in front in real time and determine whether the distance reaches a preset activation threshold; An image acquisition module 200 is connected to the distance measurement and condition judgment module 100 and is used to acquire a two-dimensional digital image of the target area when the distance reaches a preset activation threshold; A shadow analysis module 300 is connected to the image acquisition module 200 and is used to extract edge features from the two-dimensional digital image and analyze whether a cargo shadow area is formed in the front operating area through grayscale value comparison; The coordinate positioning module 400 is connected with the distance measuring and condition judging module 100 and the shadow analysis module 300 respectively, and is configured to determine the highest point coordinate, the lowest point coordinate, the leftmost point coordinate and the rightmost point coordinate of the shadow area based on the principle of perspective projection and the interval when the shadow area of the goods exists. The size calculation module 500 is connected with the coordinate positioning module 400, and is configured to calculate the height size of the goods according to the highest point coordinate and the lowest point coordinate, and calculate the width size of the goods according to the horizontal interval of the leftmost point coordinate and the rightmost point coordinate. The early warning execution module 600 is connected with the size calculation module 500, and is configured to generate a hierarchical early warning signal and control the AGV forklift to stop urgently when the height size exceeds a first safety threshold and / or the width size exceeds a second safety threshold.

[0032] Preferably, the shadow analysis module 300 comprises a pallet recognition unit 310 and a shadow verification unit 320, Figure 2 A schematic diagram of the principle of a shadow analysis module provided by way of example for the present embodiment is shown in Figure 2 The pallet recognition unit 310 is configured to identify whether a target pallet exists in a target area based on the contour features of a preset pallet. The shadow verification unit 320 is connected with the pallet recognition unit 310, and is configured to analyze whether a goods shadow area is formed in a predetermined area on the top of the target pallet through grayscale value comparison when the target pallet is identified.

[0033] In the present exemplary embodiment, the shadow analysis module 300 realizes accurate identification and anti-interference detection of the goods shadow area through the cooperative work of the pallet recognition unit 310 and the shadow verification unit 320. The pallet recognition unit 310 first performs structured analysis on the two-dimensional digital image output by the image acquisition module 200, for example, using an algorithm based on shape template matching, and compares and identifies through a pre-stored standard pallet contour library (containing the geometric features of wooden, plastic and metal pallets). The unit has a built-in adaptive threshold mechanism, for example, when the similarity of the detection area and the template contour exceeds 85%, it is determined that the target pallet exists, effectively filtering the interference of shelf columns and other similar pallet structures.

[0034] ​The shadow verification unit 320 performs secondary verification on the basis of the pallet recognition, and confirms the existence of the shadow area of the goods through gray scale gradient analysis. The unit adopts a double-threshold comparison strategy: first, the average gray scale of a predetermined area (for example, a horizontal strip 1.2 meters away from the front end of the fork) at the upper end of the pallet is calculated, and then the pixel gray scale of the area is compared with that of the background area. For example, the shadow judgment threshold is set to 70% of the background gray scale, and when the average gray scale of a continuous 3*3 pixel area is lower than the threshold, the shadow area contour tracking algorithm is started. This design effectively suppresses the interference of ambient light and still maintains the detection stability in the scene of fluctuating illumination in the warehouse.

[0035] The double unit realizes efficient operation through a pipeline processing architecture: the pallet recognition unit 310 adopts a lightweight CNN model for fast classification, the model parameter quantity is compressed to less than 500 KB, and the single-frame inference time is shortened to 8 milliseconds; the shadow verification unit 320 realizes real-time gray scale analysis through FPGA hardware acceleration, and the processing delay is controlled to be less than 2 milliseconds. This design shortens the overall processing period of the shadow analysis module 300, and meets the real-time requirements of the high-speed operation scene of the AGV forklift.

[0036] The module significantly improves the detection reliability through a double-factor verification mechanism. Experiments show that in the environment of dense shelves and uneven illumination, the pallet recognition accuracy of the module remains at a high level, and the false detection rate of the shadow area is significantly reduced. Compared with the traditional single-feature detection scheme, this design reduces the false triggering frequency of the system, and is particularly suitable for high-automation warehouse systems, provides a reliable triggering judgment basis for the ultra-high and ultra-wide detection function, and effectively supports the intelligent and safe operation of the AGV forklift.

[0037] In a preferred embodiment, referring to Figure 3 , Figure 3 A schematic diagram of a pallet recognition unit provided for the example of the present embodiment, the pallet recognition unit 310 comprises: A preprocessing subunit 311 for performing preprocessing on the two-dimensional digital image; A feature comparison subunit 312 connected with the preprocessing subunit 311, for comparing the preprocessed data with a three-dimensional contour feature database; A protrusion extraction subunit 313 for extracting protruding structures through an edge detection algorithm; A size screening subunit 314 connected with the protrusion extraction subunit 313, for screening protruding structures that meet the preset support body size specification and are not less than three; A topological verification subunit 315 connected with the size screening subunit 314, for verifying whether the relative positions of the protruding structures meet the spatial geometric constraint conditions of the tripod pallet; The output subunit 316 is configured to confirm the existence of the target tray when the size screening and / or the topological verification are satisfied.

[0038] In the present exemplary embodiment, the tray recognition unit 310 realizes accurate recognition of the target tray through a multi-level processing architecture. The preprocessing subunit 311 first performs optimization processing on the two-dimensional digital image output by the image acquisition module 200, such as eliminating sensor noise by using a Gaussian filter algorithm (filter kernel size 5x5 pixels), and then enhances the contrast between the tray edge and the background by histogram equalization to improve the gray difference of the support foot area. The feature comparison subunit 312 performs fast matching based on a three-dimensional profile feature database, such as extracting local features of the image using the SURF algorithm, and triggers depth verification when the similarity between the detection area and the pre-stored tray model exceeds 85%, effectively filtering out interference such as shelf columns.

[0039] The protrusion extraction subunit 313 accurately extracts the tray support foot profile by using an improved Canny edge detection algorithm (double threshold set to 50 / 100) combined with morphological opening operation (kernel size 3x3 pixels). The size screening subunit 314 performs geometric filtering according to the pre-set support body specifications (height 50-150 mm, width 30-80 mm), and experiments show that this design significantly reduces the false alarm rate. The topological verification subunit 315 confirms the tray type by spatial geometric constraints, such as verifying whether three support feet form an isosceles triangle distribution (bottom side length deviation ≤±10%, top angle deviation ≤±5°), and integrates perspective distortion correction function to ensure verification accuracy.

[0040] The output subunit 316 constructs a double-factor judgment mechanism: when the feature comparison similarity exceeds 85% and the topological verification is passed, or the size screening and the topological verification are satisfied at the same time, the system automatically confirms the existence of the target tray. This design significantly improves the successful recognition rate in experimental tests, and is particularly suitable for multi-tray mixed operation scenarios. Through the three-level filtering mechanism of preprocessing, feature comparison, and geometric verification, the tray recognition unit 310 realizes high-reliability detection in complex warehouse environments, provides accurate tray positioning data for subsequent shadow area analysis, and effectively supports the intelligent and safe operation of AGV forklifts.

[0041] In a preferred embodiment, referring to Figure 4 , Figure 4 Another principle schematic diagram of a shadow analysis module provided for the present embodiment is shown in the figure, and the shadow analysis module 300 comprises: The gray scale analysis unit 330 is configured to perform gray scale analysis on the preset detection area in the two-dimensional digital image, and output the shadow area to be verified when the gray scale average is lower than the preset percentage of the background area. The contour extraction unit 340 is connected with the gray scale analysis unit 330, and is configured to perform edge detection and contour extraction on the to-be-verified shadow area, and calculate a contour integrity index P e [0, 1]; The gradient calculation unit 350 is configured to extract the gray scale gradient feature of the edge pixel of the to-be-verified shadow area and calculate a gradient amplitude standard deviation G. The confidence score unit 360 is connected with the contour extraction unit 340 and the gradient calculation unit 350, and is configured to calculate a confidence score Cscore based on the contour integrity index P and the gradient amplitude standard deviation G, and the calculation formula is: Wherein, a and β are preset weight coefficients, and a + β = 1, G max is a preset maximum gradient amplitude standard deviation; The determination unit 370 is connected with the confidence score unit 360, and is configured to determine that the cargo shadow area when C score ≥ a preset score threshold; The thermal imaging verification unit 380 is connected with the determination unit 370, and is configured to start the infrared thermal imaging module for secondary verification when C score < a preset score threshold.

[0042] In the example embodiment, the shadow analysis module 300 realizes accurate determination of the cargo shadow area through a multi-modal perception fusion mechanism. The gray scale analysis unit 330 first performs regional gray scale statistics on the two-dimensional digital image output by the image acquisition module 200, for example, divides the picture into a 16x16 pixel detection grid, and marks the grid as a to-be-verified shadow area when the gray scale average of any grid is lower than 30% of the background area (the background area is automatically selected by a dynamic threshold method). This design effectively suppresses the uneven lighting interference in the warehouse scene, and still maintains the detection stability in the environment illumination fluctuation scene.

[0043] The contour extraction unit 340 adopts a sub-pixel level edge detection algorithm to perform contour analysis on the to-be-verified shadow area, for example, extracts the edge through a Canny operator, applies a Douglas-Peucker algorithm for contour simplification, and calculates the coincidence degree of the simplified contour and a standard rectangular template. When the contour integrity index P ≥ 0.85 (the P value is calculated by the intersection-over-union ratio), it is determined that the contour meets the characteristics of the pallet type cargo, effectively excluding pedestrians, shelves and other non-rectangular interference objects.

[0044] The gradient calculation unit 350 extracts the gray scale gradient feature of the edge pixel through a Sobel operator, for example, calculates the gradient amplitude in the horizontal and vertical directions, and calculates the standard deviation G. When G ≤ 15 (the preset maximum gradient amplitude standard deviation Gmax = 25), it indicates that the edge gray scale changes gently, which meets the gradual change characteristics of the shadow area, and this index can effectively identify the fuzzy edge interference such as ground stains.

[0045] The confidence score unit 360 constructs a multi-feature fusion decision model, for example, sets the weight coefficients a = 0.6, b = 0.4, and calculates the comprehensive confidence by the formula Cscore = 0.6P + 0.4(1-G / 25). When Cscore is greater than or equal to 0.75, it is directly determined as an effective shadow area. This weighting strategy balances the contribution degrees of contour integrity and edge sharpness.

[0046] The determination unit 370 integrates a double-threshold decision mechanism. When the optical detection confidence is insufficient (Cscore < 0.75), the infrared thermal imaging verification unit 380 is automatically triggered. For example, for wooden cargo, Tmin = 25°C and Tmax = 35°C are set, and the false judgment caused by black light-absorbing materials is excluded by detecting the thermal radiation characteristics of the cargo, forming a redundant verification system of light-thermal fusion. Experiments show that this design significantly improves the accuracy of shadow area recognition, especially suitable for complex scenes where high-reflective packaging goods and dark shelves coexist, and provides a reliable trigger judgment basis for ultra-high and ultra-wide detection functions.

[0047] In a preferred embodiment, the shadow analysis module 300 further comprises: An illumination sensing unit for measuring the ambient light intensity Lux value of the front work area; A dynamic threshold unit connected with the illumination sensing unit, for dynamically adjusting the gray threshold interval based on the Lux value The adjustment formula is: ; wherein b is the illumination adaptation coefficient, Lux base is the calibrated ambient light reference value; A shadow detection unit connected with the dynamic threshold unit, for detecting the shadow area based on the gray value after adjusting the threshold interval.

[0048] In this embodiment, the shadow analysis module 300 improves the reliability of cargo shadow area detection through an environment-adaptive threshold adjustment mechanism. The illumination sensing unit integrates a high-precision illumination sensor, for example, uses a TSL2591 chip to monitor the ambient light intensity of the work area in real time, with a measurement range of 0.1 lux to 220,000 lux and a resolution of 0.01 lux. After detecting the ambient light intensity Lux value, the dynamic threshold unit performs nonlinear correction according to the pre-calibration parameters, for example, sets the illumination adaptation coefficient b = 0.15 and the reference value Lux_base = 500 lux, and calculates the dynamic threshold value by the formula T_gray = T_base x [1 + 0.15 x log(Lux / 500)]. This logarithmic function design makes the threshold adjustment amplitude and the ambient light intensity show a sub-linear relationship, avoiding the detection shock caused by the sudden change of threshold in strong light environment.

[0049] In a typical application scenario, when the measured Lux = 1000, the threshold is corrected to T_gray = T_base x (1 + 0.15 x log2) ≈ 1.105T_base; when Lux = 2000, the correction is T_gray = T_base x (1 + 0.15 x log4) ≈ 1.196T_base. This hierarchical adjustment strategy makes the system control the threshold increase within 20% when the illumination intensity is doubled, effectively balancing the strong light inhibition and weak light sensitivity requirements. The shadow detection unit uses an adaptive threshold binarization algorithm, for example, combining the Otsu global threshold method with the dynamically adjusted T_gray, and determining the shadow area when the pixel gray value is lower than T_gray. Experiments show that in the day and night light difference scene of the warehouse scene, this design significantly improves the accuracy of shadow area detection and reduces the false detection rate.

[0050] The module forms an environment-adaptive shadow detection capability through the closed-loop architecture of light intensity perception-dynamic modeling-threshold optimization. Compared with the traditional fixed threshold scheme, this design effectively overcomes the detection failure problem caused by changes in illumination, and can maintain stable detection performance in strong light direct radiation (such as summer noon sunlight) or weak light compensation (such as night LED lighting) scenes. The dynamic threshold adjustment mechanism enables the system to automatically adapt to the lighting conditions of different work periods, especially suitable for intelligent warehouse scenes with multiple shifts of continuous operation, providing reliable environmental adaptability support for the ultra-high and ultra-wide detection function of AGV forklifts.

[0051] In a preferred embodiment, the system further comprises: a tilt angle sensing module for measuring the real-time pitch angle θ of the AGV forklift; a height correction module connected with the size calculation module 500 and the tilt angle sensing module, respectively, for correcting the height size of the goods based on the real-time pitch angle θ, and the correction formula is: wherein, H is the original height value, is the fork deformation compensation amount.

[0052] In this embodiment, the AGV forklift goods ultra-high and ultra-wide detection system realizes the dynamic size correction function through an integrated hardware architecture. The tilt angle sensing module uses a dual-axis MEMS sensor array, which is fixedly installed on the fork root structural member, and captures the rotation angle θ of the forklift around the horizontal axis in real time, with a sampling frequency set to 50Hz to ensure the timeliness of the attitude data. The module establishes a data link with the central processing unit through an I2C interface, and synchronously transmits attitude parameter packages containing time stamps. The height correction module as the core processing unit has a non-volatile memory pre-stored with a fork load-deformation characteristic curve, for example, when the bearing mass changes in increments of 100kg, the corresponding deformation compensation amount Linear growth characteristics with 0.8mm intervals. In the work flow, the module first receives the original height value output by the visual detection module , which is calculated by the pixel height of the goods captured by the monocular camera and the pre-calibrated camera internal parameter; then the current frame θ value of the tilt angle sensing module and the deformation compensation parameter are called to perform composite correction operation . In a specific working scenario, when θ=5° (at this time cosθ takes the value of 0.996), the visual system measures =1.5m and according to the current 200kg load, the table lookup obtains =1.6mm, the correction module calculates =1.5×0.996+0.0016≈1.4956m through the hardware floating point operation unit. Through the synergistic mechanism of sensor fusion and algorithm compensation, the device not only eliminates the posture deviation of the vehicle body by using the geometric projection principle, but also offsets the metal fatigue effect through deformation compensation, ensuring that the goods height detection value is always referenced to the horizontal reference surface when the AGV performs acceleration start-stop or slope driving in complex working conditions, providing a precision guarantee of ±2mm level for the fork lifting control, effectively reducing the collision risk of high-position access operation, and significantly improving the operation safety of warehouse logistics scene.

[0053] In a preferred embodiment, the early warning execution module 600 further comprises: A cooperative request unit for sending a detection request containing shadow area coordinates to i AGV forklifts in the surrounding when the goods width is out of limit; A data fusion unit connected with the cooperative request unit, for receiving the multi-view image data and confidence score C score,i of the goods shadow area returned by each AGV forklift in the surrounding; A three-dimensional modeling unit connected with the data fusion unit, for generating a three-dimensional point cloud model through feature point matching; A weighted calculation unit for weighted average calculation of the comprehensive goods width value according to the confidence score: ; W i The original width measurement value of each AGV forklift in the surrounding.

[0054] In this embodiment, the AGV forklift cargo overheight and overwidth detection system warning execution module 600 realizes accurate composite detection of width and height dimensions through a multi-machine cooperative architecture. When the cargo width exceeds the safety threshold, the cooperative request unit broadcasts a standardized detection protocol package to the surrounding i AGV forklifts through the vehicle-mounted wireless communication module. The protocol includes the three-dimensional space coordinates (x, y) of the cargo shadow area and the dynamic detection radius parameters. The data fusion unit receives the multi-view image data stream returned by the surrounding AGVs, and each frame of data is attached with a timestamp and camera pose information. For example, the main AGV can receive image sequences from 3 cooperative devices, each of which collects cargo side image at different pitch angles (such as 15°, 30°, 45°). The three-dimensional modeling unit uses the SIFT feature matching algorithm to perform spatial registration on multi-source images, and generates a cargo three-dimensional point cloud model containing millions of discrete points through the principle of triangulation. This model can accurately restore the spatial geometric relationship of the cargo edge profile. The weighted calculation unit has a confidence evaluation model built-in, which generates a weight coefficient C score based on the clarity, overlap, and pose accuracy of each view image. For example, when the 3 cooperative AGVs return width measurement values W1=1.78m (C score =0.95), W2=1.82m (C score =0.88), and W3=1.76m (C score =0.92), respectively, the system performs weighted fusion operation =(1.78×0.95+1.82×0.88+1.76×0.92) / (0.95+0.88+0.92)=1.79m. This device forms a detection network dynamically, eliminates monocular vision blind area using spatial geometric constraints, effectively suppresses low-quality data interference through confidence weighting mechanism, and controls measurement error within ±15mm through three-dimensional reconstruction technology, significantly improving the detection reliability of over-limit cargo in complex warehouse scenarios.

[0055] Embodiment Three Referring to Figure 5 , Figure 5 is a flowchart of an AGV forklift cargo overheight and overwidth detection method provided by an embodiment of the present application. The AGV forklift cargo overheight and overwidth detection method comprises: S100, real-time monitoring the distance between the AGV forklift and the front work area; S200, when the distance reaches a preset activation threshold, acquiring a two-dimensional digital image of the target area; S300, performing edge feature extraction on the two-dimensional digital image, and analyzing whether the front work area forms a cargo shadow area through gray value comparison; S400, if the cargo shadow area exists, determining the highest point coordinate, the lowest point coordinate, the leftmost point coordinate and the rightmost point coordinate of the cargo shadow area according to the principle of perspective projection; S500, calculating the cargo height size based on the highest point coordinate and the lowest point coordinate, and calculating the cargo width size based on the horizontal distance between the leftmost point coordinate and the rightmost point coordinate; S600, when the cargo height size exceeds the first safety threshold and / or the cargo width size exceeds the second safety threshold, generating a hierarchical early warning signal, and controlling the AGV forklift to perform an emergency stop operation.

[0056] The detection method of the embodiment is based on a dynamic triggering principle to construct an intelligent detection process. Among them The core of setting the distance monitoring and condition triggering mechanism is to solve the contradiction of energy consumption existing in the traditional multi-sensor system. When the AGV forklift is driving in an open area, the continuous operation of the detection module will cause energy waste, and the dynamic triggering mechanism enables the system to start the detection process only when the forklift approaches the potential work area, which not only ensures the timeliness of detection and matches the work process, but also avoids the resource consumption caused by invalid work of the sensor. The dual verification mechanism of gray gradient analysis and perspective projection modeling aims to solve the problem of misidentification of reflective surface and lens distortion error existing in single detection method. The hierarchical early warning design balances the continuity of work and the safety of the system, and avoids the influence of frequent false triggering on warehouse efficiency through multi-level response mechanism.

[0057] In one possible implementation manner: The distance monitoring (step S100) is realized by using a millimeter wave radar or a laser radar to monitor the distance of the front obstacle in real time in a high-frequency sampling mode (such as a 50Hz refresh rate); When the monitoring distance reaches a preset threshold (such as 1.5m), an industrial camera or a global shutter CMOS camera (resolution ≥200 million pixels) is triggered to collect a two-dimensional image of the work area and convert it into a gray image; The shadow area identification (step S300) uses an improved Canny edge detection algorithm combined with an Otsu adaptive threshold segmentation technology to determine the effective shadow area when the gray level difference of adjacent pixels exceeds a preset discrimination threshold (such as 30 level difference); The coordinate calculation (step S400) maps the feature points in the image coordinate system to the world coordinate system through a pre-calibrated perspective projection matrix, and the formula is: (Note: s is a scale factor, [X Y Z]T is the world coordinate, [u v1]T is the image homogeneous coordinate, and M is a 3×3 perspective transformation matrix); The size calculation (step S500) calculates the cargo height and width according to the extreme points of the mapped space coordinates by using the Euclidean distance formula; The early warning control (step S600) sends the hierarchical instruction to the AGV controller through the CAN bus, and the emergency stop response time is controlled within 200ms.

[0058] The embodiment realizes the following technical effects through the synergy of dynamic triggering and multi-modal verification: The interval threshold triggering mode effectively reduces the system energy consumption during unnecessary detection period, and forms an intelligent detection mode working on demand. Combining the gray gradient analysis and geometric feature verification can significantly improve the recognition reliability of the cargo shadow area, and effectively suppress the environmental light interference. The perspective projection modeling is used to eliminate the measurement error caused by the camera installation angle, and ensure the size detection accuracy. The hierarchical early warning and emergency stop linkage mechanism is constructed, a safe closed loop of detection-decision-execution is formed, and the collision accidents in the warehouse operation are effectively prevented.

[0059] Embodiment four The embodiment of the present application also provides a computer readable storage medium, the storage medium stores at least one program code, the at least one program code is loaded and executed by a processor to realize the AGV forklift cargo overheight and overwidth detection method as described in the above embodiment three. The computer program is stored on the computer readable storage medium, and the program is executed by the processor.

[0060] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment method can be implemented by a computer program to instruct related hardware, and the program can be stored in a computer readable storage medium. The program can include the processes of the above-mentioned embodiment when executed.

[0061] Embodiment five The embodiment of the present application also provides an electronic device, Figure 6 is a schematic diagram of the principle of an electronic device according to the embodiment of the present application, as Figure 6 shown, the electronic device 600 includes a memory 610, a processor 620, and a computer program 630 stored in the memory and executable on the processor, and the processor executes the computer program to realize the steps of the AGV forklift cargo overheight and overwidth detection method as described in embodiment three.

[0062] The above merely is the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation, or direct or indirect application in other related technical fields, which is made by using the content of the present application specification and drawings, is also included in the patent protection scope of the present application.

Claims

1. An AGV forklift cargo overheight and overwidth detection system, characterized in that: include: The distance measurement and condition judgment module is used to monitor the distance between the AGV forklift and the working area in front in real time and determine whether the distance reaches the preset activation threshold; An image acquisition module, connected to the distance measurement and condition judgment module, for acquiring a two-dimensional digital image of the target area when the distance reaches a preset activation threshold; a shadow analysis module, connected to the image acquisition module, for extracting edge features from the two-dimensional digital image and analyzing whether a cargo shadow area is formed in the front operating area by comparing grayscale values; a coordinate positioning module, connected to the distance measurement and condition judgment module and the shadow analysis module, respectively, for determining the coordinates of the highest point, the lowest point, the leftmost point, and the rightmost point of the shadow area based on the perspective projection principle and the distance when a shadow area of ​​the cargo exists; a size calculation module, connected to the coordinate positioning module, for calculating the height of the cargo based on the coordinates of the highest point and the lowest point, and calculating the width of the cargo based on the horizontal distance between the coordinates of the leftmost point and the rightmost point; The early warning execution module is connected to the dimension calculation module and is used to generate a graded early warning signal and control the AGV forklift to stop suddenly when the height dimension exceeds the first safety threshold and / or the width dimension exceeds the second safety threshold.

2. The system according to claim 1, wherein The shadow analysis module includes: A pallet identification unit, configured to identify whether a target pallet exists in a target area based on a preset pallet contour feature; The shadow verification unit is connected to the pallet identification unit and is used to analyze whether a cargo shadow area is formed in a predetermined area on the upper end of the target pallet through grayscale value comparison when the target pallet is identified.

3. The system according to claim 2, wherein: The pallet identification unit includes: a preprocessing subunit, configured to perform preprocessing on the two-dimensional digital image; a feature comparison subunit, connected to the preprocessing subunit, for comparing the preprocessed data with the three-dimensional contour feature database; A convex extraction subunit, used to extract convex structures through edge detection algorithm; a size screening subunit, connected to the protrusion extraction subunit, for screening at least three protrusion structures that meet the preset support body size specifications; A topology verification subunit, connected to the size screening subunit, is used to verify whether the relative positions of the protruding structures meet the spatial geometric constraints of the tripod tray; Output subunit that confirms the presence of the target pallet if size screening and / or topology verification are met.

4. The system according to claim 1, wherein: The shadow analysis module includes: A grayscale analysis unit, configured to perform grayscale analysis on a preset detection area in the two-dimensional digital image, and output a shadow area to be verified when the grayscale mean is lower than a preset percentage of the background area; A contour extraction unit, connected to the grayscale analysis unit, is used to perform edge detection and contour extraction on the shadow area to be verified, and calculate a contour integrity index P∈[0,1]; A gradient calculation unit, configured to extract the grayscale gradient features of the edge pixels of the shadow area to be verified and calculate the standard deviation G of the gradient amplitude; The confidence scoring unit is connected to the contour extraction unit and the gradient calculation unit, and is used to calculate the confidence score Cscore based on the contour integrity index P and the gradient amplitude standard deviation G. The calculation formula is: , where α and β are preset weight coefficients, and satisfy α+β=1, G max is the preset maximum gradient amplitude standard deviation; The judgment unit is connected to the confidence scoring unit and is used to score When the threshold is greater than or equal to the preset score, it is determined to be a cargo shadow area; Thermal imaging verification unit, connected to the judgment unit, used when C score <When the preset scoring threshold is reached, the infrared thermal imaging module is activated for secondary verification.

5. The system according to claim 1, wherein: The shadow analysis module also includes: A light sensor unit is used to measure the Lux value of the ambient light intensity in the front working area; A dynamic threshold unit, connected to the light sensing unit, is used to dynamically adjust the grayscale threshold interval based on the Lux value , the adjustment formula is: ; Among them, β is the light adaptation coefficient, Lux base To calibrate the ambient light reference value; The shadow detection unit is connected to the dynamic threshold unit and is used to detect the shadow area based on the grayscale value after adjusting the threshold interval.

6. The system according to claim 1, wherein: Also includes: Inclination sensor module, used to measure the real-time pitch angle θ of the AGV forklift; The height correction module is connected to the size calculation module and the tilt sensor module respectively, and is used to correct the height size of the cargo based on the real-time pitch angle θ. The correction calculation formula is: ,in, H raW is the original height value, Δh cal is the fork deformation compensation.

7. The system according to claim 4, wherein: The early warning execution module also includes: The collaborative request unit is used to send a detection request containing the coordinates of the shadow area to the surrounding i AGV forklifts when the cargo width exceeds the limit; The data fusion unit is connected to the collaborative request unit and is used to receive the multi-view image data and confidence score C of the cargo shadow area sent back by the surrounding AGV forklifts. score,i ; A three-dimensional modeling unit, connected to the data fusion unit, for generating a three-dimensional point cloud model by matching feature points; A weighted calculation unit is used to perform weighted average calculation on the comprehensive cargo width value according to the confidence score: ; W i It is the original width measurement value of each surrounding AGV forklift.

8. A method for detecting excessive height and width of cargo on an AGV forklift, characterized in that: The steps include: S100, real-time monitoring of the distance between the AGV forklift and the working area ahead; S200, when the distance reaches a preset activation threshold, acquiring a two-dimensional digital image of the target area; S300, extracting edge features from the two-dimensional digital image, and analyzing by grayscale value comparison whether a cargo shadow area is formed in the front operating area; S400: If the cargo shadow area exists, determine the coordinates of the highest point, the lowest point, the leftmost point, and the rightmost point of the cargo shadow area based on the perspective projection principle; S500: Calculate the height of the cargo based on the coordinates of the highest point and the lowest point, and calculate the width of the cargo based on the horizontal distance between the coordinates of the leftmost point and the coordinates of the rightmost point; S600: When the height dimension of the cargo exceeds a first safety threshold and / or the width dimension of the cargo exceeds a second safety threshold, a graded warning signal is generated, and the AGV forklift is controlled to perform an emergency stop operation.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for detecting excessive height and excessive width of cargo on an AGV forklift are implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the AGV forklift cargo overheight and overwidth detection method are implemented as described in any one of claims 1 to 7.

Images