A method and system for cargo inspection and supervision based on unmanned aerial vehicles (UAVs)

Abstract

This invention relates to the field of cargo management technology, and particularly to a method and system for cargo inspection and supervision based on unmanned aerial vehicles (UAVs). The method includes: scanning cargo to construct a basic cargo storage model; constructing a virtual inspection route and controlling the UAV to move along the virtual route; constructing a path control random function and controlling the UAV's flight altitude according to the path control random function to scan the cargo and obtain real-time scan data; generating a simulated path and generating simulated scan data based on the simulated path, and comparing the results to generate cargo inspection results. During the inspection process, this invention uses a UAV to randomly scan cargo along a random path, thereby obtaining scan data corresponding to each cargo model. Based on the scan data, it determines whether there are any anomalies, greatly reducing labor intensity, enabling cargo supervision over large areas, and ensuring cargo safety.

Description

Technical Field

[0001] This invention belongs to the field of cargo management technology, and in particular relates to a cargo inspection and supervision method and system based on drones. Background Technology

[0002] Cargo inspection and supervision refers to the real-time monitoring and management of goods during warehousing and transportation through a series of technical and management measures to ensure the safety, integrity, and compliance of goods. This includes using technologies such as RFID, GPS, and IoT sensors to track the location, status, and environmental conditions of goods, as well as implementing regular manual inspections and audits to prevent and detect potential risks, ensuring the smooth operation of the supply chain and the quality of goods.

[0003] In the current cargo storage management process, due to the large area of ​​the site, it is impossible to conduct manual inspections, making it difficult to ensure the safety of the goods. Summary of the Invention

[0004] The purpose of this invention is to provide a cargo inspection and supervision method based on drones, which aims to solve the problem that in the current cargo storage management process, due to the large area of ​​the site, it is impossible to conduct inspections manually, making it difficult to ensure the safety of the cargo.

[0005] This invention is implemented as follows: a method for cargo inspection and supervision based on unmanned aerial vehicles (UAVs), the method comprising:

[0006] The goods are scanned to construct a basic model of goods storage, which includes a site model and a goods model.

[0007] A virtual inspection route is constructed based on a cargo storage model, and the drone is controlled to move along the virtual inspection route.

[0008] When the drone inspects the corresponding cargo model, a path control random function is constructed, and the drone's flight altitude is controlled according to the path control random function to scan the cargo and obtain real-time scan data.

[0009] Based on the drone's flight path, a corresponding simulated path is generated in the cargo storage basic model. Simulated scan data is generated based on the simulated path, and cargo inspection results are generated by comparison.

[0010] Preferably, the step of constructing a virtual inspection route based on the cargo storage model and controlling the drone to move along the virtual inspection route specifically includes:

[0011] Based on the basic cargo storage model, a cargo storage plan is generated. The cargo storage plan is then divided into areas according to the location of the cargo model, resulting in cargo areas and free areas.

[0012] The number of pixels contained in each cargo area in the cargo storage plan is counted, and the area pixel coordinates are generated. Based on the area pixel coordinates, a function is fitted to generate a random function for the inspection order. The horizontal axis of the area pixel coordinates is the cargo area number, and the vertical axis is the number of pixels contained in the cargo area.

[0013] Import a preset sequence of natural numbers, determine the inspection order based on the calculated values, generate a virtual inspection route, and control the drone to move along the virtual inspection route.

[0014] Preferably, the step of constructing a path control random function, controlling the drone's flight altitude according to the path control random function, and scanning the cargo to obtain real-time scan data when the drone is inspecting the corresponding cargo model specifically includes:

[0015] When the drone inspects the corresponding cargo model, a set of functions is randomly selected from the pre-selected functions as the path control random function;

[0016] Using the current time value as a random variable, import it into a path-controlled random function to obtain the calculated value of the variable, and then truncate it into multiple types of strings according to a preset truncation method;

[0017] A spiral is generated based on the dimensions of the cargo model. The detection starting point and pitch are determined based on a string to obtain the scanning spiral. The UAV scans along the scanning spiral to obtain real-time scanning data.

[0018] Preferably, the steps of generating a corresponding simulated path in the cargo storage basic model based on the drone's flight path, generating simulated scan data based on the simulated path, and generating cargo inspection results by comparison specifically include:

[0019] Based on the flight path of the drone, a corresponding simulated path is generated in the cargo storage basic model, and the ranging angle of the drone at each position on the simulated path is obtained.

[0020] Calculate the distance between each point on the simulated path and the cargo model at the corresponding ranging angle to obtain simulated scan data;

[0021] The simulated scanning data is compared with the real-time scanning data one by one, the model matching rate is calculated, and the cargo inspection results are generated based on the preset threshold range.

[0022] Preferably, when the model matching rate is lower than a preset value, the corresponding cargo model is captured by an unmanned aerial vehicle (UAV) using images or videos, which are then entered into the cargo inspection results, and a warning message is issued.

[0023] Another object of the present invention is to provide a cargo inspection and monitoring system based on unmanned aerial vehicles (UAVs), the system comprising:

[0024] The model building module is used to scan the goods and build a basic model of goods storage, which includes a site model and a goods model.

[0025] The inspection route generation module is used to construct virtual inspection routes based on the cargo storage basic model and control the drone to move along the virtual inspection routes.

[0026] The real-time scanning module is used to construct a path control random function when the drone is inspecting the corresponding cargo model. The drone's flight altitude is controlled according to the path control random function to scan the cargo and obtain real-time scanning data.

[0027] The results analysis module is used to generate a corresponding simulated path in the cargo storage basic model based on the flight path of the UAV, generate simulated scanning data based on the simulated path, and generate cargo inspection results by comparison.

[0028] Preferably, the inspection route generation module includes:

[0029] The area division unit is used to generate a cargo storage plan based on the cargo storage basic model. The cargo storage plan is divided into areas according to the location of the cargo model to obtain cargo areas and free areas.

[0030] The random function generation unit is used to count the number of pixels contained in each cargo area in the cargo storage plan, generate region pixel coordinates, perform function fitting based on the region pixel coordinates, and generate a random function for inspection order. The horizontal axis of the region pixel coordinates is the number of the cargo area, and the vertical axis is the number of pixels contained in the cargo area.

[0031] The path generation unit is used to import a preset sequence of natural numbers, determine the inspection order based on the calculated values, generate a virtual inspection route, and control the drone to move along the virtual inspection route.

[0032] Preferably, the real-time scanning module includes:

[0033] The function selection unit is used to randomly select a set of functions from the pre-selected functions as the path control random function when the UAV inspects the corresponding cargo model;

[0034] The string truncation unit is used to take the current time value as a random variable, import it into a path-controlled random function to obtain the calculated value of the variable, and truncate it into multiple types of strings according to a preset truncation method;

[0035] The real-time scanning unit generates a spiral line based on the dimensions of the cargo model, determines the detection starting point and pitch based on a string, and obtains the scanning spiral line. The UAV scans along the scanning spiral line to obtain real-time scanning data.

[0036] Preferably, the result analysis module includes:

[0037] The ranging information acquisition unit is used to generate a corresponding simulated path in the cargo storage basic model based on the flight path of the UAV, and to acquire the ranging angle of the UAV at each position on the simulated path.

[0038] The simulation scanning unit is used to calculate the distance between each point on the simulation path and the cargo model at the corresponding ranging angle, thus obtaining simulation scanning data.

[0039] The inspection and judgment unit is used to compare the simulated scanning data with the real-time scanning data one by one, calculate the model matching rate, and generate the cargo inspection results based on the preset threshold range.

[0040] Preferably, when the model matching rate is lower than a preset value, the corresponding cargo model is captured by an unmanned aerial vehicle (UAV) using images or videos, which are then entered into the cargo inspection results, and a warning message is issued.

[0041] This invention provides a drone-based cargo inspection and supervision method. Based on the cargo to be supervised, a basic cargo storage model is established. In subsequent inspections, the drone randomly scans the cargo along a random path to obtain scan data corresponding to each cargo model. Based on the scan data, it is determined whether there are any anomalies. This method greatly reduces labor intensity, enables cargo supervision of large areas, and ensures cargo safety. Attached Figure Description

[0042] Figure 1 A flowchart of a cargo inspection and supervision method based on unmanned aerial vehicles (UAVs) provided for embodiments of the present invention;

[0043] Figure 2 A flowchart illustrating the steps of constructing a virtual inspection route based on a cargo storage model and controlling a drone to move along the virtual inspection route, as provided in this embodiment of the invention.

[0044] Figure 3 The flowchart of the steps provided in this embodiment of the invention is as follows: when a drone is inspecting a corresponding cargo model, a path control random function is constructed, the flight altitude of the drone is controlled according to the path control random function, the cargo is scanned, and real-time scanning data is obtained.

[0045] Figure 4The flowchart illustrates the steps of generating a corresponding simulated path in a cargo storage basic model based on the flight path of a drone, generating simulated scanning data based on the simulated path, and generating cargo inspection results by comparison, as provided in this embodiment of the invention.

[0046] Figure 5 An architecture diagram of a cargo inspection and monitoring system based on unmanned aerial vehicles (UAVs) provided in an embodiment of the present invention;

[0047] Figure 6 An architecture diagram of an inspection route generation module provided in an embodiment of the present invention;

[0048] Figure 7 An architecture diagram of a real-time scanning module provided in an embodiment of the present invention;

[0049] Figure 8 This is an architecture diagram of a result analysis module provided in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0051] like Figure 1 The diagram shows a flowchart of a cargo inspection and supervision method based on unmanned aerial vehicles (UAVs) provided by an embodiment of the present invention. The method includes:

[0052] S100, Scan the goods and construct a basic model of goods storage, which includes a site model and a goods model.

[0053] In this step, the goods are scanned. Scanning can be performed using a drone or manually. Using a drone is more efficient. Specifically, the drone is equipped with a high-precision camera, LiDAR, and GNSS positioning module. After determining the area to be scanned, the specific locations of the goods stacks are marked. By scanning each stack of goods, a corresponding model is constructed in the 3D model to obtain the basic model of the goods storage. Modeling can be done using LiDAR scanning or a visual approach, such as multi-view geometry technology, which reconstructs the 3D structure of the scene from multiple images taken from different angles. This includes steps such as feature detection and matching, camera calibration, and 3D reconstruction, for example, SfM (Structure from Motion) and MVS (Multi-View Stereo). These modeling methods are common techniques in this field and will not be elaborated further.

[0054] The S200 constructs a virtual inspection route based on a cargo storage model and controls the drone to move along the virtual inspection route.

[0055] In this step, a virtual inspection route is constructed based on the cargo storage basic model. During inspection, a drone is used to achieve the inspection effect. However, if a fixed path is used for inspection, it is easy for criminals to find the inspection pattern, allowing them to evade the drone. This invention generates region pixel coordinates based on the images corresponding to each cargo model in the cargo storage basic model. Since the stacking method of different cargoes is random, the number of pixels they occupy is also random, resulting in random region pixel coordinates. A function is fitted based on multiple sets of region pixel coordinates to obtain a set of random inspection order functions. The order of inspecting each cargo model is determined according to this random inspection order function. Since the inspection order is not fixed, the inspection time for each cargo model is also different, making it impossible for criminals to predict the inspection time and thus preventing them from evading the drone. After generating the virtual inspection route, the drone is controlled to move along the virtual inspection route and scan the cargo models during the process.

[0056] When the S300 drone is inspecting the corresponding cargo model, it constructs a path control random function, controls the drone's flight altitude according to the path control random function, scans the cargo, and obtains real-time scan data.

[0057] In this step, when the drone inspects the corresponding cargo model, a path control random function is constructed. During scanning, the drone carries a LiDAR to measure the distance between the drone and the cargo model. The path control random function is randomly selected during the scanning process and is used to control the scanning path of the cargo model. The scanning is performed in a spiral ascent, and the path control random function is used to change the pitch of the spiral ascent. Since the pitch is randomly generated and the starting position is different, the scanning path is random each time. This greatly reduces the number of sampling points while ensuring sampling effect, reducing the power consumption of the drone and improving the drone's endurance. The scanned data is saved to obtain real-time scan data.

[0058] The S400 generates a corresponding simulated path in the cargo storage basic model based on the flight path of the drone, generates simulated scanning data based on the simulated path, and generates cargo inspection results by comparison.

[0059] In this step, a corresponding simulated path is generated in the cargo storage basic model based on the drone's flight path. The drone's position is acquired in real time, and its corresponding position in the cargo storage basic model is determined based on the drone's real-time position. The parameters of the drone during the scanning process, including the scanning angle, are acquired to simulate the corresponding scanning process in the cargo storage basic model. The distance between the drone and the cargo model is calculated to obtain simulated scanning data. The simulated scanning data is compared with the real-time scanning data to determine whether there has been a change in the cargo. If a change has occurred, an anomaly is determined. The drone is used to capture images or videos of the corresponding cargo model, which are then recorded in the cargo inspection results, and an alert is issued.

[0060] like Figure 2 As shown, in a preferred embodiment of the present invention, the step of constructing a virtual inspection route based on a cargo storage model and controlling the drone to move along the virtual inspection route specifically includes:

[0061] S201, Generate a cargo storage plan based on the cargo storage basic model, and divide the cargo storage plan into regions according to the location of the cargo model to obtain cargo areas and idle areas.

[0062] In this step, a cargo storage plan is generated based on the cargo storage basic model. The cargo storage plan is a top view and is divided according to the location of the cargo. Specifically, image recognition technology is used to identify the location of the cargo and divide it into cargo areas. The part excluding the cargo areas is the free area. During scanning, the drone can travel through the free area.

[0063] S202, count the number of pixels contained in each cargo area in the cargo storage plan, generate area pixel coordinates, perform function fitting based on the area pixel coordinates, and generate a random function for inspection order. The horizontal axis of the area pixel coordinates is the cargo area number, and the vertical axis is the number of pixels contained in the cargo area.

[0064] In this step, the number of pixels contained in each cargo area within the cargo storage plan is counted. Each cargo area is then numbered, and the number can be a natural number, such as 1, 2, 3, etc. The number of pixels contained in each cargo area is then counted. For example, cargo area 1 contains 198,723 pixels, and cargo area 2 contains 284,374 pixels. Based on this, a region pixel coordinate is constructed for each cargo area, such as (1,1987). A preset number of region pixel coordinates are selected, such as five randomly selected region pixel coordinates: (1,1987), (3,2612), (5,2387), (4,1885), and (7,2081). Then, a function is fitted based on the above five region pixel coordinates to obtain a random function for the inspection order.

[0065] S203: Import a preset sequence of natural numbers, determine the inspection order based on the calculated values, generate a virtual inspection route, and control the drone to move along the virtual inspection route.

[0066] In this step, a preset sequence of natural numbers is imported and fed into the inspection order random function. The random function then generates calculated values, which are concatenated to form a string. The string is then searched according to the concatenation order to extract the order in which the corresponding numbers for each cargo area appear. If the maximum number of a cargo area does not exceed two digits, the concatenated string is split into multiple short strings consisting of two-digit decimal characters to determine the order in which each cargo area appears. The inspection order is determined based on this order. For example, if all cargo area numbers are found at the 1283rd character, the first 1283 characters are discarded, and all remaining characters in the concatenated string are retained for use in determining the inspection order in the next inspection. After determining the order in which the drone inspects each cargo area, a corresponding virtual inspection route is generated, and the drone is controlled to move along this route.

[0067] like Figure 3 As shown, in a preferred embodiment of the present invention, the step of constructing a path control random function, controlling the flight altitude of the drone according to the path control random function, and scanning the cargo to obtain real-time scan data when the drone is inspecting the corresponding cargo model specifically includes:

[0068] S301: When the UAV inspects the corresponding cargo model, it randomly selects a set of functions from the pre-selected functions as the path control random function.

[0069] In this step, when the drone is inspecting the corresponding cargo model, it queries the preset alternative function database. The alternative function database stores multiple preset random functions. When calling the function, one set of random functions is selected as the path control random function. The path control random function is used to determine the path that the drone scans for the cargo.

[0070] S302: Using the current time value as a random variable, import it into the path control random function to obtain the calculated value of the variable, and then truncate it into multiple types of strings according to the preset truncation method.

[0071] In this step, the current time value is used as a random variable to import the path control random function, thereby calculating a calculated value, which is defined as the path control calculated value. This calculated value is then divided into three parts: the first value P1, the second value P2, and the third value P3, where the third value P3 is a two-digit number.

[0072] S303 generates a spiral line based on the dimensions of the cargo model, determines the detection starting point and pitch based on the string, and obtains the scanning spiral line. The UAV scans along the scanning spiral line to obtain real-time scanning data.

[0073] In this step, a spiral is generated based on the dimensions of the cargo model. This spiral is a cylindrical spiral. Specifically, the starting position is determined, and a minimum inscribed circle is constructed based on the currently inspected cargo area. An equidistant circle is then constructed based on the minimum inscribed circle, with the radius difference between the equidistant circle and the minimum inscribed circle being R0. Taking the due north direction of the equidistant circle as the starting point, the starting point offset angle θ = (360 * P1) / (P1 + P2) is calculated. Starting from the due north direction, the spiral is offset clockwise by θ. The position of the equidistant circle at the starting point offset angle θ is the starting point of the spiral. The radius of the equidistant circle is the radius of the spiral. The pitch L = L0 * P3 / 100 is calculated, where L0 is the preset pitch. The radius of the generated spiral is equal to the radius of the equidistant circle, and the pitch of the spiral is L, resulting in a scanning spiral. The UAV flies along the scanning spiral to determine the centroid of the cargo area. During scanning, the UAV uses laser ranging, and the centroid of the cargo area is always located on the laser beam, thereby generating multiple ranging results and obtaining real-time scanning data.

[0074] like Figure 4 As shown in the preferred embodiment of the present invention, the steps of generating a corresponding simulated path in the cargo storage basic model based on the UAV flight path, generating simulated scanning data based on the simulated path, and generating cargo inspection results by comparison specifically include:

[0075] S401 generates a corresponding simulated path in the cargo storage basic model based on the UAV's flight path, and obtains the ranging angle of the UAV at each position on the simulated path.

[0076] S402, calculate the distance between each point on the simulated path and the cargo model at the corresponding ranging angle to obtain simulated scan data.

[0077] In this step, a corresponding simulated path is generated in the cargo storage basic model based on the UAV's flight path. That is, the simulated path is the same as the scanning spiral. A simulated measurement ray is generated in the cargo storage basic model. The starting point of the simulated measurement ray is the position of the UAV on the simulated path. The simulated measurement ray will pass through the centroid of the cargo area. The distance value between the simulated measurement ray and the UAV and the cargo model is counted and used as the simulated scanning data.

[0078] S403 compares the simulated scan data with the real-time scan data one by one, calculates the model matching rate, and generates the cargo inspection results based on the preset threshold range.

[0079] In this step, the simulated scan data and real-time scan data are compared one by one to determine the location of the scan point corresponding to the real-time scan data. The data that matches the simulated scan data is selected, and the distance values ​​contained in the two are compared to see if they are the same. The difference between the two is calculated. If the difference is greater than a preset value, it is determined that there is a scanning error at the scan point. The model matching rate is calculated as the ratio between the number of scan points with scanning errors and the total number of scan points. The cargo inspection results are generated based on a preset threshold range. If the model matching rate exceeds the preset threshold range, it is determined that there is an anomaly. At this time, the drone is controlled to perform image or video acquisition on the corresponding cargo.

[0080] like Figure 5 As shown, an embodiment of the present invention provides a cargo inspection and monitoring system based on unmanned aerial vehicles (UAVs). The system includes:

[0081] The model building module 100 is used to scan the goods and build a basic model of goods storage, which includes a site model and a goods model.

[0082] In this system, the model building module 100 scans the goods. Scanning can be performed using a drone or manually. Using a drone is more efficient. Specifically, the drone is equipped with a high-precision camera, LiDAR, and GNSS positioning module. After determining the area to be scanned, the specific location of the goods stacks is marked. By scanning each stack of goods, a corresponding model is built in the 3D model to obtain the basic model of the goods storage. The modeling method can be LiDAR scanning or a visual solution, such as multi-view geometry technology, which reconstructs the 3D structure of the scene from multiple images taken from different angles. This includes steps such as feature detection and matching, camera calibration, and 3D reconstruction, for example, SfM (Structure from Motion) and MVS (Multi-View Stereo). These modeling methods are common techniques in this field and will not be elaborated further.

[0083] The inspection route generation module 200 is used to construct a virtual inspection route based on the cargo storage basic model and control the drone to move along the virtual inspection route.

[0084] In this system, the inspection route generation module 200 constructs a virtual inspection route based on the cargo storage basic model. During inspection, a drone is used to achieve the inspection effect. However, if a fixed path is used for inspection, it is easy for criminals to find the inspection pattern, causing them to evade the drone based on the inspection pattern. This invention generates region pixel coordinates based on the images corresponding to each cargo model in the cargo storage basic model. Since the stacking method of different cargoes is random, the number of pixels they occupy is also random, and the resulting region pixel coordinates are also random. Based on the obtained multiple sets of region pixel coordinates, a function fitting is performed to obtain a set of random inspection order functions. The order of inspection of each cargo model is determined according to this random inspection order function. Since the inspection order is not fixed, the inspection time for each cargo model is also different, making it impossible for criminals to grasp the inspection time and thus impossible for them to evade. After generating the virtual inspection route, the drone is controlled to move along the virtual inspection route and scan the cargo models during the process.

[0085] The real-time scanning module 300 is used to construct a path control random function when the UAV is inspecting the corresponding cargo model. The UAV's flight altitude is controlled according to the path control random function to scan the cargo and obtain real-time scanning data.

[0086] In this system, the real-time scanning module 300 constructs a path control random function when the UAV inspects the corresponding cargo model. During scanning, the UAV carries a LiDAR to measure the distance between the UAV and the cargo model. The path control random function is randomly selected during the scanning process and is used to control the scanning path of the cargo model. The scanning adopts a spiral ascent method, and the path control random function is used to change the pitch of the spiral ascent. Since the pitch is randomly generated and the starting position is different, the scanning path is random each time. While ensuring the sampling effect, the number of sampling points is greatly reduced, the power consumption of the UAV is reduced, and the endurance of the UAV is improved. The scanned data is saved to obtain real-time scanning data.

[0087] The results analysis module 400 is used to generate a corresponding simulated path in the cargo storage basic model based on the flight path of the UAV, generate simulated scanning data based on the simulated path, and generate cargo inspection results by comparison.

[0088] In this system, the result analysis module 400 generates a corresponding simulated path in the cargo storage basic model based on the UAV's flight path, acquires the UAV's position in real time, determines its corresponding position in the cargo storage basic model based on the UAV's real-time position, acquires the UAV's parameters during the scanning process, including the scanning angle, and thus simulates the corresponding scanning process in the cargo storage basic model. It calculates the distance between the UAV and the cargo model to obtain simulated scanning data, compares the simulated scanning data with the real-time scanning data to determine whether there has been a change in the cargo. If a change has occurred, an anomaly is determined, and the corresponding cargo model is captured by the UAV using image or video, which is then recorded in the cargo inspection results, and an alert is issued.

[0089] like Figure 6 As shown, in a preferred embodiment of the present invention, the inspection route generation module 200 includes:

[0090] The area division unit 201 is used to generate a cargo storage plan based on the cargo storage basic model, and to divide the cargo storage plan into areas according to the location of the cargo model, so as to obtain cargo areas and idle areas.

[0091] In this module, the area division unit 201 generates a cargo storage plan based on the cargo storage basic model. The cargo storage plan is a top view and is divided according to the location of the cargo. Specifically, image recognition technology is used to identify the location of the cargo and divide it into cargo areas. The part excluding the cargo areas is the free area. During scanning, the UAV can travel through the free area.

[0092] The random function generation unit 202 is used to count the number of pixels contained in each cargo area in the cargo storage plan, generate region pixel coordinates, perform function fitting based on the region pixel coordinates, and generate a random function for inspection order. The horizontal axis of the region pixel coordinates is the cargo area number, and the vertical axis is the number of pixels contained in the cargo area.

[0093] In this module, the random function generation unit 202 counts the number of pixels contained in each cargo area within the cargo storage plan, and assigns a number to each cargo area, which can be a natural number such as 1, 2, 3, etc. Then, it counts the number of pixels contained in each cargo area; for example, cargo area 1 contains 198,723 pixels, and cargo area 2 contains 284,374 pixels. Based on this, a region pixel coordinate is constructed for each cargo area, such as (1, 1987). A preset number of region pixel coordinates are selected, such as five randomly selected region pixel coordinates: (1, 1987), (3, 2612), (5, 2387), (4, 1885), and (7, 2081). Then, a function is fitted based on these five region pixel coordinates to obtain the random function for the inspection order.

[0094] The path generation unit 203 is used to import a preset sequence of natural numbers, determine the inspection order based on the calculated value, generate a virtual inspection route, and control the drone to move along the virtual inspection route.

[0095] In this module, the path generation unit 203 imports a preset sequence of natural numbers and inputs it into the inspection order random function. The random function then generates calculated values, which are continuously concatenated to obtain a concatenated string. The string is then searched according to the concatenation order to extract the order in which the corresponding numbers for each cargo area appear. If the maximum number of a cargo area does not exceed two digits, the concatenated string is split into multiple short strings consisting of two-digit decimal characters to determine the order in which each cargo area appears. The inspection order is determined based on the order in which each cargo area appears. For example, if all cargo area numbers are found at the 1283rd character, the first 1283 characters are discarded, and all remaining characters in the concatenated string are retained for use in determining the inspection order during the next inspection. After determining the order in which the drone inspects each cargo area, a corresponding virtual inspection route is generated, and the drone is controlled to move along the virtual inspection route.

[0096] like Figure 7 As shown, in a preferred embodiment of the present invention, the real-time scanning module 300 includes:

[0097] The function selection unit 301 is used to randomly select a set of functions from the pre-selected functions as the path control random function when the UAV inspects the corresponding cargo model.

[0098] In this module, when the UAV inspects the corresponding cargo model, the function selection unit 301 queries the preset alternative function database. The alternative function database stores multiple preset random functions. When calling, it selects one set of random functions as the path control random function. The path control random function is used to determine the path for the UAV to scan the cargo.

[0099] The string truncation unit 302 is used to take the current time value as a random variable, import it into the path control random function to obtain the calculated value of the variable, and truncate it into multiple types of strings according to the preset truncation method.

[0100] In this module, the string truncation unit 302 uses the current time value as a random variable to import the path control random function, thereby calculating a calculated value, which is defined as the path control calculated value. The calculated value is then divided into three parts: the first value P1, the second value P2, and the third value P3, where the third value P3 is a two-digit number.

[0101] The real-time scanning unit 303 is used to generate a spiral line according to the size of the cargo model, determine the detection starting point and pitch based on the string, and obtain the scanning spiral line. The UAV scans along the scanning spiral line to obtain real-time scanning data.

[0102] In this module, the real-time scanning unit 303 generates a spiral line based on the dimensions of the cargo model. This spiral line is a cylindrical spiral line. Specifically, the starting position is determined, and a minimum inscribed circle is constructed based on the currently inspected cargo area. An equidistant circle is then constructed based on the minimum inscribed circle, with the radius difference between the equidistant circle and the minimum inscribed circle being R0. Taking the due north direction of the equidistant circle as the starting point, the starting point offset angle θ = (360 * P1) / (P1 + P2) is calculated. Starting from the due north direction, the spiral line is offset clockwise by θ. The position of the equidistant circle at the starting point offset angle θ is the starting point of the spiral line, and the radius of the equidistant circle is the radius of the spiral line. The pitch L = L0 * P3 / 100 is calculated, where L0 is the preset pitch. The radius of the generated spiral line is equal to the radius of the equidistant circle, and the pitch of the spiral line is L, thus obtaining the scanning spiral line. The UAV flies along the scanning spiral line to determine the centroid of the cargo area. During scanning, the UAV uses laser ranging, and the centroid of the cargo area is always located on the laser ray, thereby generating multiple ranging results and obtaining real-time scanning data.

[0103] like Figure 8 As shown, in a preferred embodiment of the present invention, the result analysis module 400 includes:

[0104] The ranging information acquisition unit 401 is used to generate a corresponding simulated path in the cargo storage basic model based on the flight path of the UAV, and to acquire the ranging angle of the UAV at each position on the simulated path.

[0105] The simulation scanning unit 402 is used to calculate the distance between each point on the simulation path and the cargo model at the corresponding ranging angle, and obtain simulation scanning data.

[0106] In this module, the ranging information acquisition unit 401 generates a corresponding simulated path in the cargo storage basic model based on the flight path of the UAV. That is, the simulated path is the same as the scanning spiral. A simulated measurement ray is generated in the cargo storage basic model. The starting point of the simulated measurement ray is the position of the UAV on the simulated path. The simulated measurement ray will pass through the centroid of the cargo area. The distance value between the simulated measurement ray and the UAV and the cargo model is counted and used as the simulated scanning data.

[0107] The inspection judgment unit 403 is used to compare the simulated scanning data with the real-time scanning data one by one, calculate the model matching rate, and generate the cargo inspection results based on the preset threshold range.

[0108] In this module, the inspection and judgment unit 403 compares the simulated scanning data with the real-time scanning data one by one to determine the location of the scanning point corresponding to the real-time scanning data. It selects matching data from the simulated scanning data and compares whether the distance values ​​contained in the two are the same. It calculates the difference between the two. If the difference is greater than a preset value, it is determined that there is a scanning error at the scanning point. The model matching rate is calculated as the ratio between the number of scanning points with scanning errors and the total number of scanning points. Based on a preset threshold range, the cargo inspection result is generated. If the model matching rate exceeds the preset threshold range, it is determined that there is an anomaly. At this time, the drone is controlled to perform image or video acquisition on the corresponding cargo.

[0109] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for cargo inspection and supervision based on unmanned aerial vehicles (UAVs), characterized in that, The method includes: The goods are scanned to construct a basic model of goods storage, which includes a site model and a goods model. A virtual inspection route is constructed based on a cargo storage model, and the drone is controlled to move along the virtual inspection route. When the drone inspects the corresponding cargo model, a path control random function is constructed, and the drone's flight altitude is controlled according to the path control random function to scan the cargo and obtain real-time scan data. Based on the flight path of the drone, a corresponding simulated path is generated in the basic model of cargo storage. Simulated scanning data is generated according to the simulated path, and cargo inspection results are generated by comparison. The steps involved in constructing a virtual inspection route based on a cargo storage model and controlling a drone to move along that route include: Based on the basic cargo storage model, a cargo storage plan is generated. The cargo storage plan is then divided into areas according to the location of the cargo model, resulting in cargo areas and free areas. The number of pixels contained in each cargo area in the cargo storage plan is counted, and the area pixel coordinates are generated. Based on the area pixel coordinates, a function is fitted to generate a random function for the inspection order. The horizontal axis of the area pixel coordinates is the cargo area number, and the vertical axis is the number of pixels contained in the cargo area. Import a preset sequence of natural numbers, determine the inspection order based on the calculated values, generate a virtual inspection route, and control the drone to move along the virtual inspection route; The steps involved in generating a simulated path based on the drone's flight path in the cargo storage basic model, generating simulated scan data based on the simulated path, and generating cargo inspection results through comparison include: Based on the flight path of the drone, a corresponding simulated path is generated in the cargo storage basic model, and the ranging angle of the drone at each position on the simulated path is obtained. Calculate the distance between each point on the simulated path and the cargo model at the corresponding ranging angle to obtain simulated scan data; The simulated scanning data is compared with the real-time scanning data one by one, the model matching rate is calculated, and the cargo inspection results are generated based on the preset threshold range.

2. The unmanned aerial vehicle (UAV)-based cargo inspection and supervision method according to claim 1, characterized in that, The step of constructing a path control random function, controlling the drone's flight altitude according to the path control random function, and scanning the cargo to obtain real-time scan data when the drone is inspecting the corresponding cargo model specifically includes: When the drone inspects the corresponding cargo model, a set of functions is randomly selected from the pre-selected functions as the path control random function; Using the current time value as a random variable, import it into a path-controlled random function to obtain the calculated value of the variable, and then truncate it into multiple types of strings according to a preset truncation method; A spiral is generated based on the dimensions of the cargo model. The detection starting point and pitch are determined based on a string to obtain the scanning spiral. The UAV scans along the scanning spiral to obtain real-time scanning data.

3. The unmanned aerial vehicle (UAV)-based cargo inspection and supervision method according to claim 1, characterized in that, When the model matching rate is lower than the preset value, the drone will collect images or videos of the corresponding cargo model, record them in the cargo inspection results, and issue a warning message.

4. A cargo inspection and monitoring system based on unmanned aerial vehicles (UAVs), characterized in that, The system includes: The model building module is used to scan the goods and build a basic model of goods storage, which includes a site model and a goods model. The inspection route generation module is used to construct virtual inspection routes based on the cargo storage basic model and control the drone to move along the virtual inspection routes. The real-time scanning module is used to construct a path control random function when the drone is inspecting the corresponding cargo model. The drone's flight altitude is controlled according to the path control random function to scan the cargo and obtain real-time scanning data. The results analysis module is used to generate corresponding simulated paths in the cargo storage basic model based on the flight path of the UAV, generate simulated scanning data based on the simulated paths, and generate cargo inspection results by comparison. The inspection route generation module includes: The area division unit is used to generate a cargo storage plan based on the cargo storage basic model. The cargo storage plan is divided into areas according to the location of the cargo model to obtain cargo areas and free areas. The random function generation unit is used to count the number of pixels contained in each cargo area in the cargo storage plan, generate region pixel coordinates, perform function fitting based on the region pixel coordinates, and generate a random function for inspection order. The horizontal axis of the region pixel coordinates is the number of the cargo area, and the vertical axis is the number of pixels contained in the cargo area. The path generation unit is used to import a preset sequence of natural numbers, determine the inspection order based on the calculated values, generate a virtual inspection route, and control the drone to move along the virtual inspection route. The results analysis module includes: The ranging information acquisition unit is used to generate a corresponding simulated path in the cargo storage basic model based on the flight path of the UAV, and to acquire the ranging angle of the UAV at each position on the simulated path. The simulation scanning unit is used to calculate the distance between each point on the simulation path and the cargo model at the corresponding ranging angle, thus obtaining simulation scanning data. The inspection and judgment unit is used to compare the simulated scanning data with the real-time scanning data one by one, calculate the model matching rate, and generate the cargo inspection results based on the preset threshold range.

5. The cargo inspection and monitoring system based on unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The real-time scanning module includes: The function selection unit is used to randomly select a set of functions from the pre-selected functions as the path control random function when the UAV inspects the corresponding cargo model; The string truncation unit is used to take the current time value as a random variable, import it into a path-controlled random function to obtain the calculated value of the variable, and truncate it into multiple types of strings according to a preset truncation method; The real-time scanning unit generates a spiral line based on the dimensions of the cargo model, determines the detection starting point and pitch based on a string, and obtains the scanning spiral line. The UAV scans along the scanning spiral line to obtain real-time scanning data.

6. The cargo inspection and monitoring system based on unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, When the model matching rate is lower than the preset value, the drone will collect images or videos of the corresponding cargo model, record them in the cargo inspection results, and issue a warning message.

Images