METHOD AND SYSTEM FOR LOAD DETECTION IN AN INSTALLATION CONVEYOR

Description

[0001] The invention relates to a method for load detection in a forklift truck, wherein the load is detected by means of at least one optical sensor and is recognized by evaluating the sensor data of the sensor using artificial intelligence in at least one data processing unit, and to a system for carrying out the method.

[0002] Industrial trucks are used for transporting and / or handling loads. One example of this is order picking. Order picking refers to all methods for assembling specific load objects, especially individual items such as packages, from a provided assortment in warehouses. This assembly is based on orders, such as customer orders or production orders. Recently, automated systems have been increasingly used for this purpose. In this process, individual items are picked up manually by a picker or by autonomously operated transport vehicles from a source location, especially a source load carrier, such as a source pallet, and placed onto a destination load carrier, such as a destination pallet, transported on the vehicle. Mobile picking robots, which can independently pick up the individual items using robotic arms, are particularly common in this process.These mobile picking robots automatically receive order data such as order number, storage location coordinates, quantity, and weight of the items from a central computer. They can, for example, navigate to specific storage locations in racking systems and retrieve the desired item using a handling device, particularly a gripping system with a picking tool. The picking tool can be a gripper, such as an adhesion gripper or a vacuum gripper.

[0003] The vehicles used for order picking are usually self-driving or are driven by an autonomous vehicle (automated guided vehicle, or AGV) and usually have a lifting mast or other lifting device for vertical positioning of the gripper in the rack and on the target container.

[0004] The loads to be transported are generally not standardized. Therefore, the loads usually differ in the following factors: weight, weight distribution, height, width, depth, number of individual load objects, three-dimensional distribution on the loading platform of the forklift truck, center of gravity of the load in the application axis (vertical Z-axis), horizontal distance of the load's center of gravity, etc.

[0005] The factors mentioned are dynamic in nature. Especially during order picking, the characteristics of a load are constantly changing. This creates problems for autonomous guided vehicles (AGVs) when handling loads, because the aforementioned factors determine whether an AGV can even pick up a load. A potential load overhang, where the load extends beyond the load carrier's contour, plays a particularly crucial role, as load carriers with an unknown overhang cannot be handled easily. For example, it is impossible to determine a sufficiently large shelf space within a rack, nor can safe transport be guaranteed, since the overhanging load could indirectly widen the AGV's footprint, making personnel protection in accordance with applicable standards impossible.

[0006] Furthermore, these factors related to the transported load play a role in the performance of AGVs. Speed, turning angle, and safety aspects are all designed based on the load's characteristics. If these factors are unknown, one must assume the worst-case scenario.

[0007] There is currently no known solution for determining all factors of a load to be transported.

[0008] To ensure that the load does not protrude beyond the load carrier, separate stationary contour inspection stations are often used. These are the first stations approached by the forklift transporting the load after goods have arrived, in order to guarantee further safe transport of the load. However, this means a complex additional transport process for each individual load carrier.

[0009] Overall, it can be stated that an overhanging load poses a complex problem in any automated warehouse. Expensive load contour check stations must be set up and accessed with each new load before automated handling of the load by a forklift can be safely ensured. If such contour check stations are not available, the corresponding check must be carried out manually.

[0010] The weight and weight distribution of a load can be determined by additional sensors, especially weight sensors.

[0011] Some of the other load factors cannot be determined using simple sensors.

[0012] For safety reasons, a binary load is often assumed. This means that there is either a load on the forklift or there isn't. However, this severely limits the safety and performance of the forklift.

[0013] However, even this limited performance of binary load detection can only be achieved with extensive cabling directly at the loading platform of the forklift. Cables, however, always pose a problem for forklifts with moving platforms, especially autonomous mobile platform vehicles with lifting and lowering platforms for load transport. On the one hand, cable length can be problematic due to data transmission requirements. On the other hand, cables on moving parts, such as a lifting platform, always present a fire hazard, which can be caused by cable crushing.

[0014] Autonomous industrial trucks often handle incoming goods by picking up pallets. However, this is only possible if all three dimensions of the load are known: width, height, and depth. Otherwise, the pallet will not be picked up.

[0015] A major problem in order picking is the manual or automated picking of incorrect loads. The resulting costs are enormous.

[0016] A generic method for load detection in a forklift truck is known from WO 2020 / 215772 A1.

[0017] DE 10 2020 123 381 A1 discloses a position and location estimation system that estimates the position and location of a pallet, which is a freight handling target, with respect to a forklift truck with a pair of forks.

[0018] From EP 3 378 825 A1, a forklift truck with a flying object is known.

[0019] The present invention is based on the objective of designing a method for load detection in a forklift truck and a system for carrying out the method in such a way that reliable load detection is ensured even for complexly structured loads transported by a forklift truck.

[0020] This problem is solved according to the invention by determining the load center of gravity by evaluating the sensor data in the data processing unit.

[0021] A camera is most suitable as the optical sensor. The camera can be designed as a compact camera module that offers flexible deployment. For example, mobile phone cameras already provide sufficient quality.

[0022] The camera provides the necessary sensor data for evaluation in the data processing unit. The evaluation of the sensor data in the data processing unit is advantageously carried out using an imaging technique. Artificial intelligence methods are applied in this process.

[0023] Artificial intelligence (AI) enables the identification, evaluation, and classification of objects, such as load objects. Small, powerful computers can now be built from a hardware environment specialized for AI applications.

[0024] The optical sensor can be carried by the industrial truck itself. Preferably, a sensor mounted on a lifting mechanism of the industrial truck, for example, at the top of a lifting mast, is used. By mounting the optical sensor on the lifting mechanism, it is possible for the sensor to detect the industrial truck and the load, for example, on a loading platform of the industrial truck, from a higher position.

[0025] Additionally or alternatively, an optical sensor located outside the industrial truck can also be used.

[0026] The optical sensor can be mounted on an infrastructure element within the operational environment of the forklift truck. For example, the sensor can be positioned on a mast at a greater height to ensure accurate detection of the forklift truck and its load.

[0027] Preferably, a movable optical sensor is used that can be moved around the forklift and the load, thus capturing the forklift and the load from different directions. This facilitates the evaluation of the sensor data and load detection in the data processing unit. For this purpose, a sensor carried by an operator can be used, such as a body camera, head camera, smart glasses, or virtual reality headset.

[0028] According to a particularly preferred embodiment of the invention, the sensor is carried by a drone. Drones are already known for other applications. For several years, drones have been increasingly used in various sectors. They are popular both as a hobby among private individuals and in professional settings. Their flight capability, high-performance cameras, and connectivity with other computers and smartphones are advantageous in these applications. Drones are also beginning to be used in logistics, for example, for warehouse inventory, or their future use is being considered, such as for package delivery. Industrial drones with highly advanced technology are already available; these utilize artificial intelligence and enable, among other things, sensor fusion (e.g., lidar, camera, and radar), environmental mapping, and self-localization.

[0029] The stabilization of a drone's hovering flight has advanced to the point where load object detection by a camera mounted on the drone is easily possible. Cameras integrated into the drones themselves can also be used. The payload of a drone can currently be up to 15 kg, allowing an AI-capable computer and an additional transmitter / receiver unit to be easily carried.

[0030] Using a drone allows the onboard optical sensor to flexibly and quickly detect loads from any angle (360°). Another advantage of a drone is that it enables the detection of loads from multiple forklifts without additional effort.

[0031] Preferably, the data processing unit is also carried on the drone. Drones are readily capable of carrying compact AI modules for evaluating sensor data using artificial intelligence. This allows both the optical detection of the load and the forklift using the optical sensor and the evaluation of the sensor data using artificial intelligence to be performed on board the drone. In this case, the transmission of sensor data to an external data processing unit is unnecessary.

[0032] The data processing unit can also be carried by the forklift itself. This is particularly advantageous when the optical sensor is mounted on the forklift.

[0033] Another option is to operate the data processing unit in a stationary position. In this case, it is advisable to transmit the sensor data from the sensor carried by the drone and / or the forklift to the stationary data processing unit via a wireless data connection.

[0034] It can also be advantageous to transmit sensor data from the sensor carried on the drone to the data processing unit carried on the forklift via a wireless data connection. Conversely, it can be advantageous to transmit sensor data from the sensor carried on the forklift to the data processing unit carried on the drone via a wireless data connection.

[0035] A further development of the inventive concept provides for the use of at least two data processing units that exchange information via a wireless data connection. In this way, several data processing units can be networked together.

[0036] In an advanced stage of development, it is conceivable, for example, that several stationary data processing units, several data processing units carried by drones, and several data processing units carried by industrial trucks are provided, all exchanging information with each other via wireless data connections. In combination with multiple optical sensors, some of which can be stationary and some carried by the drones and industrial trucks, and which transmit their sensor data to their respective assigned data processing units, the maximum possible networking can be achieved.

[0037] A single drone can detect the loads of multiple industrial trucks. For this purpose, the loads of at least two industrial trucks are preferably detected by the sensor carried on the drone.

[0038] The load of a forklift can also be detected by several drones. For this purpose, the load of the forklift or the loads of at least two forklifts are advantageously detected by sensors carried by at least two drones.

[0039] According to a preferred embodiment of the invention, it is provided that the industrial truck and the associated load are detected by means of the sensor and that, by evaluating the sensor data in the data processing unit, a comparison of load data of the detected load with load data previously stored in the data processing unit for the industrial truck is carried out, wherein, in the event of a deviation in the load data, the load data for the industrial truck are updated.

[0040] For example, the camera of a drone, or another camera not attached to the forklift, or a camera directly attached to the forklift, captures the forklift and its load. The data processing unit associated with the camera compares the load data of the specific forklift with the load data previously stored in the data processing unit for that forklift. If these differ, the data processing unit updates the load data for the forklift.

[0041] Another possibility is that the drone receives the instruction from the forklift truck beforehand to check and update the load from the forklift truck.

[0042] In particular, the following load detection criteria can be met according to the invention: 1. Dimensions of the load (height, width, depth):Individual load components can be identified by evaluating the sensor data in the data processing unit, for example, using imaging techniques and artificial intelligence, and their dimensions (height, width, depth) can be determined. From this, the overall size (height, width, depth) of the load on the forklift can be calculated. Using a drone offers the advantage that it can fly a full 360 degrees around the load, allowing the optical sensor to capture the load from all sides. 2. Number of individual load objects: Individual load objects on the forklift truck can be identified by evaluating the sensor data in the data processing unit, and their number can be determined. 3. Weight and weight distribution of the load:Markers, such as QR codes, barcodes, or Aruco markers, on individual load objects can be read by evaluating the sensor data in the data processing unit, and the read information, especially weight information, can be assigned to the respective load object. The weight of individual load objects can be recorded by corresponding markers on the load objects, which are read by the sensor, and assigned to the corresponding load object. Furthermore, a load object can be recognized and classified based on its outer packaging, and its weight can then be retrieved from a database. Another possibility is to assign the weight to the load object during loading using weight sensors. This also provides clear information about the weight distribution of the load. 4. Three-dimensional distribution of the load on the industrial truck:A three-dimensional load distribution on the loading platform of a forklift truck can be determined by evaluating the sensor data in the data processing unit. Based on the dimensions of the individual load objects and the known dimensions of the forklift truck's loading platform, a specific load distribution on the loading platform is determined. 5. Load center of gravity:The center of gravity of a load on the loading platform of a forklift truck is determined by evaluating the sensor data in the data processing unit. For example, the sensor of a drone can detect the forklift truck and its load. There are two ways to detect the load's center of gravity. One method is to use the shape of the entire load to determine its vertical center of gravity, or one can first detect the geometry and position of each individual load item by analyzing the sensor data using artificial intelligence. Additionally, the weight of each load item can be transmitted via a marker, such as a QR code, Aruco marker, barcode, etc. By combining all this information, a more precise center of gravity for the entire load can be determined.The center of gravity is particularly important for autonomous platform vehicles because it allows for continuous adjustment of speed and performance. Otherwise, the worst-case scenario must be assumed. 6. Load center distance: Especially when picking up a load, the load center distance—for example, the distance of the load center from the fork carriage of a load fork—is crucial for permitting load picking. Without this information, an autonomous vehicle cannot pick up a load. The load center distance can also be determined by evaluating the sensor data in the data processing unit. 7. Quality control and pick confirmation during order picking:The data processing unit can compare the load data of the detected load with the load data of the load specified for the forklift truck in the order. If the load data matches, the forklift truck is released. If there is a discrepancy, an error message and / or a return order for the forklift truck is generated. This solves the problem of incorrectly picking loads. The weight and individual dimensions of each load, and, if applicable, of the vehicle, are typically stored in a cloud, a warehouse management system, or similar databases. The data processing unit can retrieve this information at any time. The database information about the load could also be transferred to the data processing unit at the start of the order, eliminating the need for communication during the order placement process. Quality control takes place directly during picking.For this purpose, the weight, dimensions, any markers on the currently picked load, its position, etc., are determined by evaluating the sensor data using artificial intelligence. The identified load is then compared with the information in the forklift, such as the load data of the loads to be picked. If the load has been picked correctly, the forklift receives a pick confirmation. For manually operated vehicles, this might be indicated by a green light or a release to proceed. For autonomous vehicles, it might be a release to move. If the picked load is identified as incorrect for the forklift, an error message is generated. For manually operated vehicles, this error can be signaled via a display or a signal to the order picker. For autonomous vehicles, the vehicle can be instructed to return the last load.This is possible because both the position of the load object and the location from which it was picked are known. This allows for direct comparison during order processing to ensure the order is being completed correctly. At the end, the identified and verified load objects can be compared again with the order list and confirmed to the inventory management system. This comparison with the inventory management system can also be used to automatically perform an inventory count. Picked load objects are selectively removed from the overall inventory list, ensuring that the current stock level is always visible. Of course, to record the stock level, incoming goods must also be checked. Here, too, the invention can be used to record the goods and load objects and add them to the inventory management system.

[0043] The invention further relates to a system for carrying out the method with at least one industrial truck and a load for the industrial truck.

[0044] In the system, the problem set out in the invention is solved by providing at least one data processing unit operating with artificial intelligence and operatively connected to at least one optical sensor, in particular a camera, which is configured to detect the load by evaluating the sensor data.

[0045] In an advantageous embodiment of the invention, the sensor is attached to the industrial truck, in particular to a lifting device of the industrial truck.

[0046] According to another, particularly preferred, embodiment of the invention, the sensor is attached to a flying drone.

[0047] Sensors can be attached to both the forklift and the drone.

[0048] Additionally or alternatively, at least one sensor can be permanently attached to an infrastructure element.

[0049] For the evaluation of the sensor data, the respective sensor is in operative communication with a data processing unit, which is preferably also housed in the drone.

[0050] Alternatively, the data processing unit can also be housed in the industrial truck.

[0051] Data processing units can be housed in both the forklift and the drone.

[0052] Additionally or alternatively, a data processing unit can also be housed in a stationary infrastructure facility.

[0053] If multiple data processing facilities are provided, each of these is operatively connected to a data sending and data receiving facility which is designed to exchange information with at least one other data processing facility.

[0054] The data processing unit of the drone is preferably in operative communication with a data transmission and data reception device, which is designed to exchange information with at least one other drone and / or the industrial truck.

[0055] The invention offers a number of advantages: Quality control and pick confirmation reduce operating costs during the order picking process. Furthermore, it enables fully automated load handling. The costs associated with stationary load contour check stations, which must be visited with each new load, are eliminated. Manual intervention by employees is no longer necessary. Load handling by autonomous vehicles is easily accomplished. The functionality of the autonomous operation of the forklift is not restricted. Sensor data can be processed directly in the drone. This eliminates the need for cabling, thus preventing fire hazards from moving parts. In addition, gradual speed adjustments of the forklift, especially an autonomous one, result in increased efficiency.The error rate due to incorrectly picked load objects can be significantly reduced.

[0056] Further advantages and details of the invention are explained in more detail with reference to the exemplary embodiments shown in the schematic figures. These show Figure 1 shows a system with several industrial trucks and a drone, Figure 2 shows a schematic diagram for recognizing the number and dimensions of load objects, Figure 3 shows a schematic diagram for recognizing the three-dimensional distribution of load objects, Figure 4 shows a schematic diagram for weight recognition, Figure 5 shows a schematic diagram for recognizing the load center of gravity, and Figure 6 shows a representation of the load center of gravity distance.

[0057] In the Figure 1A system 1 comprising several industrial trucks 2, 3, 4 and a drone 5 is shown. The drone 5 has an optical sensor S configured as a camera and a data processing unit D that uses artificial intelligence to evaluate the sensor data from sensor S. The data processing unit D is connected to a data transmission and reception device E. The camera is preferably movably mounted on the drone 5.

[0058] The industrial truck 2 is designed as a low-level order picker with a load fork G on which load objects O of a load L are placed. The industrial truck 2 can also be equipped with an optical sensor S, designed as a camera, and a data processing unit D, which uses artificial intelligence to evaluate the sensor data from sensor S. The data processing unit D can be connected to a data transmission and data reception device E.

[0059] The industrial truck 4 has a lifting device H, for example a lifting mast, for handling loads. In this case, the optical sensor S, designed as a camera, is mounted at the upper end of the lifting device H. This ensures a better overview for the optical sensor S. A data processing unit D is also provided, which uses artificial intelligence to evaluate the sensor data from sensor S. The data processing unit D can be connected to a data transmission and data reception device E.

[0060] The industrial truck 3 is designed as an autonomous vehicle (AGV) with a lifting platform P on which load objects O of a load L are placed. The industrial truck 3 can be equipped with an optical sensor S, designed as a camera, and a data processing unit D, which uses artificial intelligence to evaluate the sensor data from sensor S. The data processing unit D can be connected to a data transmission and data reception device E.

[0061] The load detection of the load L located on the industrial trucks 2, 3, 4 is achieved using imaging techniques. Optical sensors S and data processing units D are provided for this purpose. The drone 5 has the advantage of being able to fly around the load L and thus quickly detect its position at 360°. A further advantage of the drone 5 is its ability to detect the loads L of multiple industrial trucks 2, 3, 4.

[0062] The optical sensor S of the drone 5 provides the necessary sensor data for evaluation in the data processing unit D. The evaluation of the sensor data from sensor S in the data processing unit D is performed using an imaging technique. Artificial intelligence methods are applied in this process. By evaluating the sensor data from sensor S, the dimensions of the load objects O, their arrangement, and the number of individual load objects can be determined. The weights and weight distribution of the load objects O can also be determined by reading markers M applied to the load objects O. From this, the center of gravity LS of the load L can also be determined.

[0063] The vehicle-integrated sensors S of the industrial trucks 2, 3, 4 can also detect the loads L being lifted and the surroundings of the industrial trucks 2, 3, 4. This sensor data is evaluated in the vehicle-integrated data processing units D.

[0064] In order to exchange the information available in the drone's own data processing unit D of the drone 5 and obtained by evaluating the sensor data of the drone's own sensor S with the industrial trucks 2, 3, 4 and, if necessary, to compare it with the information available in the vehicle's own data processing units D and obtained by evaluating the sensor data of the vehicle's own sensors S, both the drone 5 and the industrial trucks 2, 3, 4 have data transmitting and data receiving devices E.

[0065] In the Figure 1Load detection of the load L on one or more industrial trucks 2, 3, 4 is preferably carried out using the drone 5 and its camera. This can be done – as shown below – using the following example: Figures 2 to 6 The load L is described as having its dimensions, the number of individual load objects O1, O2, O3, the weight and weight distribution of the load L, the three-dimensional distribution of the load L on the industrial truck 2, 3, 4, and the load center of gravity as well as the load center of gravity distance of the load L. These load L factors can be transmitted from the drone 5 to the corresponding industrial truck 2, 3, 4 via the data transmission and data reception devices E.

[0066] As in the Figure 2As shown, imaging techniques and artificial intelligence can be used to identify the individual load objects O, O1, O2, O3 of a load L and determine their dimensions, such as height, width, and depth, as well as the number of load objects O1, O2, O3. A single load object O can be classified in step A and located in step B. When multiple load objects O1, O2, O3 are present, the individual load objects O1, O2, O3 can first be detected in step C, and their dimensions, of which the Figure 2 Only the heights h1, h2, h3 are given, and the overall height h of the load L can be determined. From the heights h1, h2, h3 of the load objects O1, O2, O3, a total height h of the load L is obtained (step D). Similarly, from the other determined dimensions (width, depth), a total dimension (width, depth) of the load L is then obtained.

[0067] The dimensions of the load objects O, O1, O2, O3 can be detected, for example, during the loading of the forklift truck 2, 3, 4.

[0068] As in the Figure 3 As shown, relocated load objects O2 can also be measured by first detecting the adjacent load object O1 and determining its depth T1, and then adding the protruding part of load object O2 to this measurement, resulting in the depth T2 of load object O2. The drone 5 then checks whether load objects O1 and O2 have a flush rear end.

[0069] Based on the dimensions of the individual load objects O1, O2 and the known dimensions of the, in the Figure 2Based on the loading area of ​​the forklift truck 2, 3, 4 (or 3 not shown), a specific three-dimensional distribution of the load L on the loading area of ​​the forklift truck 2, 3, 4, which is also detected by the sensor S, is determined. This allows free areas to be used for order picking by the forklift truck 2, 3, 4.

[0070] The Figure 4 Figure 1 shows a schematic diagram for weight detection. The weight of individual load objects O can be transmitted to the drone 5 via markers M, for example QR codes, barcodes, Aruco markers, etc., by the drone 5's sensor S detecting the corresponding marker M and the drone 5's data processing unit D evaluating the sensor data from sensor S to read the marker M and assign this information to the load object O.

[0071] If additional sensors, such as weight sensors, are used to assign the weight to the load object during loading, the weight distribution of the load can also be determined.

[0072] In the Figure 5Figure 1 shows a schematic diagram for detecting the center of gravity LS of the load L. First, the load objects O1, O2, O3 are detected by the sensor S of the drone 5. The geometry and position of each individual load object O1, O2, O3 are then determined by evaluating the sensor signals from sensor S in the data processing unit D of the drone 5 using artificial intelligence. Additionally, the weight of each load object O1, O2, O3 can be transmitted to the drone 5 via a corresponding marker M, such as a QR code, Aruco marker, barcode, etc. Using all this information, a precise center of gravity LS of the entire load L is determined, specifically the vertical height of the center of gravity LS and the horizontal distance between the center of gravity and the load L.

[0073] Especially for the autonomous industrial truck 3 with the lifting platform, the load center of gravity, in particular the vertical height of the load center of gravity LS, of the load L is important because this allows the speed and performance of the autonomous industrial truck 3 to be continuously adjusted.

[0074] The Figure 6 This diagram shows a load-bearing capacity diagram illustrating the horizontal distance between load centers. The load weights are plotted on the vertical axis y, while the load center distance is plotted on the horizontal axis x. The load-bearing capacity diagram schematically depicts the relationships on, for example, a load fork G of a forklift truck 2, 3, 4. The x-axis corresponds to the load fork G and the y-axis to the fork back GR. If the load L is positioned on the load fork G of the forklift truck 2, 3, 4, then the load center distance is the distance of the load center LS from the fork back GR.

[0075] For an autonomous industrial truck in particular, the load center distance L is crucial for permitting load picking. Without this information, load picking by an autonomous vehicle is not possible.

Claims

1. Method for detecting a load in an industrial truck (2, 3, 4), wherein the load (L) is captured by means of at least one optical sensor (S) and is detected by evaluating the sensor data from the sensor (S) by means of artificial intelligence in at least one data processing unit (D), characterized in that a load centre of gravity (LS) of the load (L) is determined by evaluating the sensor data from the sensor (S) in the data processing unit (D).

2. Method according to Claim 1, characterized in that a camera is used as the optical sensor (S).

3. Method according to Claim 1 or 2, characterized in that the sensor (S) is carried by the industrial truck (2, 3, 4).

4. Method according to Claim 3, characterized in that a sensor (S) arranged on a lifting device (H) of the industrial truck (2, 3, 4) is used as the sensor (S).

5. Method according to one of Claims 1 to 4, characterized in that a sensor (S) arranged outside the industrial truck (2, 3, 4) is used as the sensor (S).

6. Method according to Claim 5, characterized in that the sensor (S) is carried by an aerial drone (5).

7. Method according to Claim 6, characterized in that the data processing unit (D) is carried by the aerial drone (5).

8. Method according to one of Claims 1 to 7, characterized in that the data processing unit (D) is carried by the industrial truck (2, 3, 4).

9. Method according to one of Claims 1 to 8, characterized in that the data processing unit (D) is operated in a stationary manner.

10. Method according to one of Claims 1 to 9, characterized in that the sensor data from the sensor (S) are transmitted to the data processing unit (D) via a wireless data connection.

11. Method according to one of Claims 1 to 10, characterized in that at least two data processing units (D) are used and exchange information via a wireless data connection.

12. Method according to one of Claims 1 to 11, characterized in that the sensor data from the sensor (S) are evaluated in the data processing unit (D) by means of an imaging method.

13. Method according to one of Claims 1 to 12, characterized in that the industrial truck (2, 3, 4) and the associated load (L) are captured by means of the sensor (S) and load data relating to the detected load (L) are compared with load data previously stored in the data processing unit (D) for the industrial truck (2, 3, 4) by evaluating the sensor data from the sensor (S) in the data processing unit (D), wherein the load data for the industrial truck (2, 3, 4) are updated in the event of a deviation of the load data.

14. Method according to one of Claims 7 to 13, characterized in that the loads (L) of at least two industrial trucks (2, 3, 4) are captured by the sensor (S) carried by the aerial drone (5).

15. Method according to one of claims 7 to 14, characterized in that the load (L) of the industrial truck (2, 3, 4) or the loads (L) of at least two industrial trucks (2, 3, 4) is / are captured by the sensors (S) carried by at least two aerial drones (5).

16. Method according to one of Claims 1 to 15, characterized in that individual load objects (O, O1, O2, O3) of the load (L) are detected by evaluating the sensor data from the sensor (S) in the data processing unit (D) and their dimensions are determined and a total size of the load (L) is calculated therefrom.

17. Method according to one of Claims 1 to 16, characterized in that individual load objects (O, O1, O2, O3) of the load (L) are detected by evaluating the sensor data from the sensor (S) in the data processing unit (D) and their number is determined.

18. Method according to one of Claims 1 to 17, characterized in that markers (M) of individual load objects (O, 01, O2, O3) of the load (L) are read out by evaluating the sensor data from the sensor (S) in the data processing unit (D) and read information, in particular weight information, is assigned to the respective load object (O, O1, O2, 03).

19. Method according to one of Claims 1 to 18, characterized in that a three-dimensional distribution of the load (L) on a loading surface of the industrial truck (2, 3, 4) is determined by evaluating the sensor data from the sensor (S) in the data processing unit (D).

20. Method according to one of Claims 1 to 19, characterized in that load data relating to the detected load (L) are compared with load data relating to the load (L) intended for the industrial truck (2, 3, 4) according to the order in the data processing unit (D), wherein a clearance for the industrial truck (2, 3, 4) is granted if the load data match and an error message and / or a return order for the industrial truck (2, 3, 4) is issued in the event of a deviation.

21. System (1) for carrying out the method according to one of Claims 1 to 20, having at least one industrial truck (2, 3, 4) and a load (L) for the industrial truck (2, 3, 4), characterized in that at least one data processing unit (D) working with artificial intelligence and operatively connected to at least one optical sensor (S), in particular a camera, is provided and is configured to detect the load (L) by evaluating the sensor data from the sensor (S).

22. System (1) according to Claim 21, characterized in that the sensor (S) is mounted on the industrial truck (2, 3, 4), in particular on a lifting device (H) of the industrial truck (2, 3, 4).

23. System (1) according to Claim 21 or 22, characterized in that the sensor (S) is mounted on an aerial drone (5).

24. System (1) according to Claim 23, characterized in that the data processing unit (D) is accommodated in the aerial drone (5).

25. System (1) according to Claim 24, characterized in that the data processing unit (D) is operatively connected to a data transmitting and data receiving device (E) which is designed to exchange information with at least one further aerial drone (5) and / or the industrial truck (2, 3, 4).