Industrial truck with an evaluation unit and a camera, logistics equipment and methods for operating such an industrial truck

Abstract

A forklift truck 2 with an evaluation unit 4 and a camera 6, logistics equipment 24, and a method for operating a forklift truck 2. The camera 6 is mounted on or in the forklift truck 2 such that at least one vehicle component of the forklift truck 2 is at least partially within the camera 6's field of view. The camera 6 is data-connected to the evaluation unit 4 and configured to transmit image data to the evaluation unit 4. The evaluation unit 4 is configured to evaluate image information contained in the image data, determine characteristics of the vehicle component, and, based on these characteristics, perform a condition classification of the vehicle component or the camera 6.

Description

[0001] The invention relates to an industrial truck with an evaluation unit and at least one camera. Furthermore, the invention relates to a logistics facility with a control computer and a plurality of such industrial trucks. The invention also relates to a method for operating an industrial truck with an evaluation unit and at least one camera.

[0002] Industrial trucks often include camera systems, for example digital camera systems, which use one or more cameras to depict the surroundings of the industrial truck or to support load handling.

[0003] Proper operation of an industrial truck includes regular inspections to ensure that the truck is in perfect working order. This inspection is performed, for example, by a technician during maintenance. Furthermore, an operator often performs a so-called pre-operation check, a brief inspection before the truck is put into operation.

[0004] It is an object of the invention to provide an industrial truck, a logistics device and a method for operating an industrial truck, wherein the operational safety is to be improved.

[0005] The object is achieved by an industrial truck with an evaluation unit and at least one camera which is attached to or in the industrial truck in such a way that at least one vehicle component of the industrial truck is located at least partially in the field of view of the camera, wherein the camera is coupled to the evaluation unit in terms of data technology and is designed to transmit image data to the evaluation unit, and wherein the evaluation unit is designed to evaluate image information present in the image data and to determine features of the vehicle component and to carry out a condition classification of the vehicle component or the camera on the basis of these features of the vehicle component.

[0006] The design of such an industrial truck is based on the following technical considerations: The camera systems commonly found on industrial trucks, which are used, for example, for environmental monitoring, not only record the surroundings of the industrial truck, but in many cases also vehicle components of the industrial truck or parts of these vehicle components. The recorded image information can be evaluated to draw conclusions about the condition of the industrial truck. In particular, the condition of the vehicle component or components that are at least partially imaged can be assessed. This evaluation is far superior to regular inspections during maintenance or a brief check before the industrial truck is put into operation (pre-op check). In many cases, minor damage to components of the industrial truck is overlooked during conventional inspections.Furthermore, it is not only possible to detect damage to the components of the industrial truck, but also to the cameras or the camera itself, for example, incorrect camera positioning. Traditionally conducted inspections often overlook damage that is not immediately visible to the service technician or the operator of the industrial truck and is therefore not visible, such as damage to the mast. The industrial truck is then put into operation anyway, which can lead to operational hazards. Violent damage, caused, for example, by a collision between a vehicle component, such as the mast, and fixed structural elements of a logistics facility, such as supports or gates, is not uncommon in the operation of industrial trucks.Even much less severe collisions can, for example, lead to a shift in the position of a vehicle component, the consequences of which can be safety-relevant. For example, a misalignment of one or more cameras can impair or even endanger the proper operation of the industrial truck. This is advantageously avoided in the industrial truck according to aspects of the invention.

[0007] In the context of the present description, a vehicle component is a component or an assembly. In particular, the vehicle component is a movable or movable vehicle component. A movable vehicle component is, in particular, a driven or drivable vehicle component, for example the lifting mast of an industrial truck or a load fork. Features of the vehicle component include, for example, its position, size, and / or its spatial or visual appearance. From the latter, it can be concluded, for example, whether the vehicle component is damaged or intact, or whether the vehicle component is missing. The industrial truck is, in particular, a forklift truck, furthermore, in particular, a reach truck.

[0008] The industrial truck is advantageously developed in that the evaluation unit has a database comprising first features which appear in the image data when the vehicle component or the camera is in a normal state, and second features which appear in the image data when the vehicle component or the camera is in a deviating state. In other words, the image data has first features when the industrial truck is in proper condition. For example, vehicle components can be seen in this image data at certain intended positions and have a certain proper appearance. The evaluation unit is configured to classify the condition of the vehicle component or the camera as "condition: normal" or "condition: deviating" based on a comparison of the image information with the first and second features.For example, if a particular vehicle component is identifiably in its intended location, it would be assigned the status "normal." However, if the vehicle component is displaced or visibly damaged, it would be assigned the status "deviating."

[0009] Alternatively, the evaluation unit comprises a pre-trained neural network which is configured to receive the image data as input data and to output a condition classification of the vehicle component as “condition: normal” or “condition: deviant”.

[0010] According to further embodiments, further state classifications can be added to the "normal" and "deviant" states. For example, a further state classification "defective" can be provided. This can also be provided as a subclassification of the "deviant" state.

[0011] The above-mentioned evaluation of the image data is carried out, for example, using conventional image processing. The features contained in the database are defined, for example, based on sample images that represent a normal or abnormal condition of the vehicle component. Information about the geometry, dimensions, or position of the vehicle component can be stored as features. A classification can be performed by comparing the captured image data with the aforementioned features, if necessary, with the addition of additional information about the geometric shape of the vehicle component.

[0012] Alternatively, or possibly in addition to this classic image analysis, image evaluation is performed using a pre-trained neural network. This network is pre-trained using training data so that, based on the image data as input, the neural network outputs the states "normal" or "deviant" with a high degree of probability.

[0013] The neural network can be trained with real image data or with synthetic image data. In the latter case, image data is simulated that contains anomalies that the neural network is supposed to detect as "deviant." One advantage of training the neural network with simulated data is that anomalies with high variance can be modeled and trained. It is also possible to train a specific state of the industrial truck as "normal." The neural network can therefore, for example, be trained as a supervised neural network, meaning that both the "normal" and "deviant" states are trained. The neural network can also be trained as an unsupervised neural network, meaning that only the "normal" states are trained. All states that are not detected as "normal" are classified as "deviant."Training of the neural network can also continue during ongoing operation. This can occur, in particular, if a condition reported as an anomaly, i.e., a condition identified as "deviating," is classified as non-critical by an operator or service technician. In such a case, the detected condition is classified as "normal," and regular operation of the industrial truck in this configuration is henceforth considered the normal condition.

[0014] According to a further embodiment, it is provided that the evaluation unit is further configured to output a message comprising the condition classification, in particular to transmit it to a control computer of a logistics facility in which the industrial truck can be operated.

[0015] To transmit the message from the industrial truck to the control computer, the industrial truck and the control computer maintain a suitable data connection, for example, a partially wireless data connection. For example, the industrial truck communicates wirelessly with an access point, from which a wired data connection to a control computer of the logistics facility is available. The message is transmitted, for example, in real time or near real time. The control computer of the logistics facility is advantageously enabled to provide a user with a detailed and up-to-date picture of the status of the industrial trucks operated in the logistics facility.

[0016] According to a further embodiment, the industrial truck is further developed in that the evaluation unit is configured to perform a condition classification of the vehicle component based on the features of the vehicle component. The industrial truck further comprises a reach mast, a load handling device, and a mast advance device, wherein the load handling device is movable along the reach mast in a lifting direction and the mast advance device is configured to move the reach mast forwards and backwards transversely to the lifting direction, and wherein the camera is mounted on or in the industrial truck such that the reach mast, as a vehicle component, is located at least partially in the camera's field of view, and wherein the evaluation unit is configured to determine a reach position of the reach mast as a condition classification.

[0017] The evaluation of the mast reach path in industrial trucks is traditionally performed using dedicated sensors or by estimating it based on one or more parameters. Often, knowing the exact mast reach position is not safety-relevant, but purely a convenience feature. Nevertheless, it may be desirable to know the exact reach position of the reach mast, for example, to implement a gentle end stop for the mast reach. If the reach position of the reach mast is determined based on the image data captured by the camera, additional sensors can be omitted, which represents a cost advantage.

[0018] According to one embodiment, a control unit is provided within the industrial truck, which can be implemented, for example, as part of the vehicle control system. Based on a value of the reach position of the reach mast, it causes a gentle end stop of the reach mast. For example, a corresponding reduction in the speed of a hydraulic pump driving the mast extension can be provided. Likewise, a suitable valve control can be provided, which controls the flow of hydraulic fluid to or from a hydraulic cylinder of the mast extension.

[0019] The traditionally used additional sensors result in additional manufacturing costs for the industrial truck. Estimating based on operating data alone may not be accurate enough, so that, despite the appropriate soft stop function, harder end stops or unnecessarily abrupt braking of the mast extension frequently occur. All of these disadvantages are advantageously avoided with the industrial truck according to the aforementioned design. Furthermore, it is advantageous that a digital camera system already present on the industrial truck can be used for this function, for example, a camera system that is otherwise used for environmental monitoring.

[0020] According to one embodiment, the evaluation unit is further configured to determine, as a feature of the reach mast, a size of an image of at least a portion of the reach mast depicted at least in part in the image data, and to use this feature to determine a reach position of the reach mast as a condition classification. For example, the size of a mast head of the reach mast can be determined, and the size of this area in the image data can be used to infer a reach position of the reach mast.

[0021] According to further advantageous embodiments, the industrial truck is further developed in that information on a geometry of the reach mast and / or reference data on a plurality of reach positions are stored in the evaluation unit and the evaluation unit is set up to determine the reach position by evaluating the size of the image of at least a part of the reach mast shown in the image data, taking into account the information on the geometry of the reach mast, and / or to determine the reach position by comparing the image information with the reference data, in particular taking additional account of the information on the geometry of the reach mast.

[0022] According to a further embodiment, the industrial truck is further developed in that the reach mast has a machine-readable identification feature, in particular a QR code and / or an ArUco marker, as a feature. The evaluation unit can be configured to determine the size of an image of the machine-readable identification feature in the image data and, taking into account information on a geometry and / or size of the machine-readable identification feature, to determine the reach position of the reach mast. Alternatively or additionally, the evaluation unit can be configured to determine the reach position of the reach mast by comparing the image information with reference data, in particular, additionally taking into account information on the geometry and / or size of the machine-readable identification feature.

[0023] Furthermore, it can be provided in particular that the evaluation unit is designed to identify the mast in the image data on the basis of the machine-readable identification feature.

[0024] The position of the mast can, for example, be directly calculated from the known size and / or geometry of the machine-readable identification feature, e.g. a QR tag, using the known camera parameters and the known imaging geometry.

[0025] Alternatively, according to a further embodiment, the industrial truck can be further developed in that the evaluation unit comprises a pre-trained neural network which is configured to receive the image data as input data and to output a push position of the reach mast.

[0026] If conventional image processing is used, for example, the edges of a reach boom head can be located in the image data. The current position of the reach boom can be determined by comparing it with reference data. A neural network can, for example, be calibrated or trained on the two end positions of the reach boom. Furthermore, the neural network can be used in combination with conventional image analysis. For example, the reach boom or the parts of the reach boom present in the image data can be identified in the image data using a neural network. Further processing then takes place based on conventional image processing. The output of such a neural network results, for example, in a rectangle in the camera image that corresponds to the image of the mast head and whose size correlates with the position of the mast extension.It can also be provided that the pre-trained neural network provides the reach position of the reach mast as an output value. Such a neural network can be trained with synthetic data, for example, based on a simulated industrial truck.

[0027] According to a further advantageous embodiment, the industrial truck is further developed in that the evaluation unit is configured to determine a plurality of push positions of the reach mast for successive points in time and further comprises a control unit which is configured to evaluate the plurality of push positions in a time-dependent manner with regard to a vibration of the reach mast, wherein the control unit is further configured, when a vibration of the reach mast is present, a) to output a warning message and / or b) to output control commands to a mast feed and / or travel control which are suitable for actively counteracting the vibration of the reach mast.

[0028] In industrial trucks, reach masts often swing, especially when loads are raised, in conjunction with movements of the truck or the mast extension. Traditional solutions rely on the use of acceleration sensors in combination with active control, for example of the drive or mast extension. However, these solutions incur high additional costs, including the components required and their installation. For example, the cable routing to the components on the mast required in such systems is complex to implement. Furthermore, the sensors used to detect mast vibrations only serve a single intended purpose and cannot be used for other functions of the industrial truck. Traditional systems for detecting mast vibrations are therefore expensive and complex.

[0029] The industrial truck according to the aforementioned embodiment advantageously uses sensors, namely at least one camera, which can also be used for other purposes, such as environmental monitoring. The installation of additional sensors can advantageously be omitted. To record the mast flexion, values ​​are obtained from the image data that allow conclusions to be drawn about a periodically changing distance between the reach mast and the chassis of the industrial truck, for example, the driver's cab. For this purpose, a camera can be mounted, for example, on the overhead guard in the direction of the forks, in order to capture, for example, a mast head or part of the mast of the reach mast and thus detect the position of the reach mast.

[0030] To compensate for vibrations, an active control system is advantageously provided, which is coupled, for example, to a mast feed and / or drive control system of the industrial truck. This active control system or control unit generates suitable control signals from the determined mast vibration values, which actively counteract the vibration of the reach boom, for example, by generating targeted counter-rotating movements of the mast feed drive and / or the drive drive based on the control signals.

[0031] The object is further achieved by a logistics facility with a control computer and a plurality of industrial trucks, each of which comprises an evaluation unit configured to output a message comprising the condition classification and transmit it to a control computer of the logistics facility. The industrial trucks operated in the logistics facility can also be configured according to one or more of the aforementioned embodiments.

[0032] Advantageously, in such a logistics facility, the control computer always has an exact picture of the safety-relevant status of the industrial trucks operated in the logistics facility.

[0033] The object is further achieved by a method for operating an industrial truck with an evaluation unit and at least one camera which is attached to or in the industrial truck in such a way that at least one vehicle component of the industrial truck is located at least partially in the image field of the camera, wherein the camera is data-technically coupled to the evaluation unit, comprising the following steps: transmitting image data from the camera to the evaluation unit, evaluating the image information present in the image data in the evaluation unit and determining features of the vehicle component and carrying out the condition classification of the vehicle component or the camera based on these features.

[0034] The same or similar advantages apply to the method of operating the industrial truck as those already mentioned with regard to the industrial truck itself, so repetition will be avoided.

[0035] The method is further advantageously developed in that the evaluation unit has a database comprising first features which appear in the image data when the vehicle component or the camera is in a normal state, and second features which appear in the image data when the vehicle component or the camera is in a deviating state, wherein the evaluation unit carries out a state classification of the vehicle component or the camera as "state: normal" or "state: deviating" based on a comparison of the image information with the first and second features, or the evaluation unit comprises a pre-trained neural network which receives the image data as input data and outputs a state classification of the vehicle component or the camera as "state: normal" or "state: deviating".

[0036] In particular, it is further provided that the evaluation unit also outputs a message comprising the condition classification and transmits it in particular to a control computer of a logistics facility in which the industrial truck is operated.

[0037] According to a further advantageous embodiment, the method is further developed in that a condition classification of the vehicle component is carried out on the industrial truck based on the features and the industrial truck further comprises a reach mast, a load handling device and a mast feed device, wherein the load handling device is movable along the reach mast in a lifting direction and the mast feed device moves the reach mast back and forth transversely to the lifting direction, and wherein the camera is attached to or in the industrial truck in such a way that the reach mast as a vehicle component is located at least partially in the image field of the camera, and wherein the evaluation unit determines a feed position of the reach mast as a condition classification.

[0038] Furthermore, it is provided in particular that a size of an image of at least a part of the reach mast shown at least in sections in the image data is determined as a feature of the reach mast and that the push position of the reach mast is determined on the basis of this feature.

[0039] Furthermore, it is provided in particular that information on a geometry of the reach mast and / or reference data on a plurality of push positions are stored in the evaluation unit and the evaluation unit evaluates the size of the image of at least part of the reach mast shown in the image data taking into account the information on the geometry of the reach mast and determines the push position and / or compares the image information with the reference data, in particular taking into account the information on the geometry of the reach mast, and determines the push position therefrom or the evaluation unit comprises a pre-trained neural network which receives image data as input data and outputs a push position of the reach mast.

[0040] It is particularly provided that the method is further developed in that the reach mast has a machine-readable identification feature, in particular a QR code and / or an ArUco marker, as a feature, and the evaluation unit determines a size of an image of the machine-readable identification feature in the image data and, taking into account information on a geometry and / or size of the machine-readable identification feature, determines the pushing position of the reach mast and / or determines the pushing position of the reach mast by comparing the image information with reference data, in particular taking into account the information on the geometry and / or size of the machine-readable identification feature.

[0041] Finally, the method is advantageously further developed in that the evaluation unit determines a plurality of push positions of the reach mast for successive points in time and the industrial truck further comprises a control unit which evaluates the plurality of push positions in a time-dependent manner with regard to a vibration of the reach mast, wherein the control unit further, when a vibration of the reach mast is present, a) issues a warning message and / or b) issues control commands to a mast feed and / or travel control which actively counteract the vibration of the reach mast.

[0042] Further features of the invention will become apparent from the description of embodiments of the invention together with the claims and the accompanying drawings. Embodiments of the invention may incorporate individual features or a combination of several features.

[0043] Within the scope of the invention, features marked with “in particular” or “preferably” are to be understood as optional features.

[0044] The invention is described below, without limiting the general inventive concept, using exemplary embodiments with reference to the drawings, whereby express reference is made to the drawings for all details of the invention not explained in more detail in the text. They show: Fig. 1 a simplified representation of an industrial truck with a reach mast and a camera, Fig. 2 a schematic representation of the image captured by the camera in Fig. 1 captured image data, Fig. 3 that from Fig. 1 known industrial truck with extended reach mast, Fig. 4 a schematic representation of the image captured by the camera in Fig. 3 captured image data, Fig. 5 a simplified representation of an industrial truck in which vibrations of the reach mast are detected Fig. 6a) and b) schematic representations of image data showing a “normal” state ( Fig. 6a) and a “deviant” condition ( Fig. 6b) show, Fig. 7 a schematic representation of a logistics facility with a control computer and a plurality of industrial trucks.

[0045] In the drawings, identical or similar elements and / or parts are provided with the same reference numbers, so that a repeated presentation is omitted.

[0046] Fig. 1 shows a simplified schematic view of an industrial truck 2 with an evaluation unit 4 and a camera 6. A single camera 6 is shown as an example. It can also be a camera system with multiple cameras 6. In particular, it is a camera system that is also used for other purposes, for example to monitor the surroundings of the industrial truck 2. Furthermore, the camera 6 is attached, for example, to the overhead guard of the industrial truck 2. According to further exemplary embodiments, the camera 6 is attached, for example, to the load handling device 8, for example to a load fork. For example, the camera 6 can be a camera that is mounted on the back of the fork or else a fork tine camera. In the following, reference will be made to a camera 6 solely for reasons of readability. The camera 6 is suitably linked to the evaluation unit 4 in terms of data technology.For this purpose, data transmission lines or wireless data transmission links not shown in the figure are provided.

[0047] The camera 6 is mounted on the industrial truck 2 such that at least one vehicle component of the industrial truck 2 is located at least partially within the field of view of the camera 6. In the illustrated embodiment, the vehicle component is the reach mast 10, which is captured in the region of its upper end, also referred to as the mast head 11. Image data captured by the camera 6 is transmitted to the evaluation unit 4 via the aforementioned data transmission paths. The evaluation unit 4 is configured to evaluate image information contained in the image data and to determine features of the vehicle component.

[0048] Fig. 2 shows an example of the Fig. 1. The upper end of the reach boom 10, comprising the mast head 11, is visible. As a feature of this exemplary captured vehicle component, namely the mast head 11 of the reach boom 10, its size is determined in the displayed camera image 12. This can be done, for example, by identifying the edges 14 of the mast head 11 or another part of the reach boom 10 in the image data and calculating a size, for example, of the resulting rectangle. With knowledge of the geometry of the reach boom 10, a position of the reach boom 10 along the travel path 16 can be deduced from the size of the rectangle (resulting from the image of the mast head 11).Thus, the evaluation unit 4 performs a condition classification of the vehicle component based on the characteristics of the vehicle component, i.e., the size of the image of the mast head 11, by assigning the reach mast 10 a specific position along the travel path 16. In other words, the pushing position of the reach mast 10 can be determined based on the camera image 12.

[0049] The reach mast 10 is moved along a travel path 16 by means of a mast feed device 18. The mast feed device 18 is, for example, a hydraulic unit configured and designed for this purpose. The load-handling device 8 is movable along the reach mast 10 in a lifting direction 20. For this purpose, the industrial truck 2 is also provided with a suitable hydraulic unit, which is not shown. The mast feed device 18 is configured to move the reach mast 10 back and forth along the travel path 16 transversely to the lifting direction 20.

[0050] Fig. 3 shows this in principle already Fig. 1 known industrial truck 2, in this case however with extended reach mast 10. For example, in Fig. 1 the reach mast 10 in a first or retracted end position and in the illustration in Fig. 3 in an extended or opposite end position.

[0051] In the situation that Fig. 3 schematically, the camera 6 records a camera image 12 as shown schematically in Fig. 4. Again, the edges 14 of an upper end of the reach mast 10, i.e., the mast head 11, can be identified in the image data, and the size of the rectangle enclosed by these edges 14 can be calculated. Based on the size of the area determined in this way, the reach position of the reach mast 10 can be determined as a condition classification.

[0052] As an alternative to determining the thrust position based on the imaged size of the mast head 11 and knowledge of the geometry of the thrust mast 10, a comparison can also be made with a plurality of reference images stored in the evaluation unit 4 as reference data. The evaluation unit 4 can determine the thrust position by comparing the captured image information with this reference data, possibly also taking into account the information on the geometry of the thrust mast 10 also stored in the evaluation unit 4.

[0053] According to a further embodiment, the reach boom 10 comprises a machine-readable identification feature, for example a QR code and / or an ArUco marker, furthermore, for example, on the boom head 11. The evaluation unit 4 can identify the boom head based on the machine-readable identification feature in the image data and subsequently carry out one of the analyses explained.

[0054] It is also provided that the evaluation unit 4 analyzes the machine-readable identification feature itself, for example, by determining the size of an image of the machine-readable identification feature in the image data. Taking into account information about the geometry and / or size of the machine-readable identification feature, the push position of the reach mast 10 can be determined. This can also be done by comparing the image information with reference data, in particular by additionally taking into account the information about the geometry and / or size of the machine-readable identification feature.

[0055] According to a further embodiment, a pre-trained neural network is present in the evaluation unit 4. This neural network is trained to provide a push position of the reach boom 10 as output data based on the image data captured by the camera 6 as input data. In this way, the push position of the reach boom 10 can also be determined based on the image data captured by the camera 6.

[0056] According to a further embodiment, the evaluation unit 4 is configured to determine a plurality of thrust positions of the reach boom 10 for consecutive points in time. For example, the position of the reach boom 10 can be detected at consecutive points in time, for example, several times per second. The evaluation unit 4 is further configured to infer a vibration of the reach boom 10 from this plurality of thrust positions, which were recorded at different points in time. This situation illustrates Fig. 5, wherein the vibration of the reach mast 10 is represented by way of example by a double arrow. Vibration of the reach mast 10 is generally undesirable. The evaluation unit 4 can further be configured to output a warning message in such a case. This warning message can be output, for example, on a display device of the industrial truck 2. Furthermore, the evaluation unit 4 can be configured to output control commands to a mast feed and / or travel control 22, wherein the mast feed and / or travel control 22 is configured and designed to counteract the vibration of the reach mast 10 by means of corresponding control commands, which optionally ensure a travel movement of the industrial truck 2 itself or a movement of the reach mast 10. In this way, the vibration of the reach mast 10 can be actively dampened.

[0057] According to further exemplary embodiments, the industrial truck 2 can also be configured such that its evaluation unit 4 comprises a database in which first features and second features are stored. The first features occur in the image data captured by the camera 6 when the vehicle component is in a normal state. The second features occur in a state of the vehicle component that deviates from the normal state. This state shall also be referred to as a "deviating state."

[0058] Fig. 6a) and Fig. 6b) each show a camera image 12 in which an upper part of the reach boom 10 is visible in the area of ​​the boom head 11. The camera image 12 in Fig. 6a) shows a normal state, the camera image in Fig. 6b) shows a different situation. The upper part of the reach mast 10, as shown schematically and simplified in Fig. 6b), has been displaced, for example, by a violent event, such as a collision. Likewise, the camera 6 may also have been displaced, for example, by external influences, resulting in a deviation in the image data. The evaluation unit 4 is configured to classify the condition of the vehicle component as "condition: normal" or "condition: abnormal" based on a comparison of the image information with the first and second features. If the "abnormal" condition is detected, a warning message may be issued, for example. This warning message is displayed, for example, on a display unit of the industrial truck 2 or transmitted to a control computer of a logistics facility 24 in which the industrial truck 2 is operated.

[0059] Alternatively, a pre-trained neural network is present in the evaluation unit 4. This network is capable of receiving the image data from camera 6 as input data and outputting a condition classification of the vehicle component as "condition: normal" or "condition: abnormal." In this way, an abnormal vehicle condition can also be detected.

[0060] Fig.Figure 7 shows a schematic and simplified representation of a logistics facility 24 in which a plurality of industrial trucks 2 are operated. The logistics facility 24 comprises a control computer 26, which communicates with the industrial trucks 2, for example, via wireless data connections 28. As described above, the messages comprising the condition classification, i.e., in particular, warning messages of the "deviating" condition, are transmitted from the industrial trucks 2 to the control computer 26, so that the control computer always has a precise picture of the condition of the industrial truck 2 operated in the logistics facility 24.

[0061] All mentioned features, including those revealed solely in the drawings as well as individual features disclosed in combination with other features, are considered essential to the invention, both individually and in combination. Embodiments according to the invention may be fulfilled by individual features or a combination of several features. List of reference symbols 2 industrial trucks 4 Evaluation unit 6 Camera 8 load handling equipment 10 reach boom 11 Masthead 12 Camera image 14 edge 16 travel path 18 Mast feed device 20 lifting device 22 Mast reach and / or travel control 24 Logistics facility 26 tax calculators

Claims

[1] Industrial truck (2) with an evaluation unit (4) and at least one camera (6) which is attached to or in the industrial truck (2) in such a way that at least one vehicle component of the industrial truck (2) is located at least partially in the image field of the camera (6), wherein the camera (6) is data-technically coupled to the evaluation unit (4) and is designed to transmit image data to the evaluation unit (4), and wherein the evaluation unit (4) is designed to evaluate image information present in the image data and to determine features of the vehicle component and to carry out a condition classification of the vehicle component or of the camera (6) on the basis of these features of the vehicle component. [2] Industrial truck (2) according to claim 1, wherein the evaluation unit (4) comprises a database comprising first features which appear in the image data when the vehicle component or the camera (6) is in a normal state, and second features which appear in the image data when the vehicle component or the camera (6) is in a deviating state, wherein the evaluation unit (4) is configured to carry out a state classification of the vehicle component or the camera (6) as “state: normal” or “state: deviating” based on a comparison of the image information with the first and second features. or the evaluation unit (4) comprises a pre-trained neural network which is configured to receive the image data as input data and to output a state classification of the vehicle component or the camera (6) as “state: normal” or “state: deviant”. [3] Industrial truck (2) according to claim 2, wherein the evaluation unit (4) is further configured to output a message comprising the condition classification and, in particular, to transmit it to a control computer (26) of a logistics facility (24) in which the industrial truck (2) is operable. [4] Industrial truck (2) according to claim 1, wherein the evaluation unit (4) is configured to carry out a condition classification of the vehicle component based on the features of the vehicle component, further comprising a reach mast (10), a load-handling means (8) and a mast feed device (18), wherein the load-handling means (8) is movable along the reach mast (10) in a lifting direction (20) and the mast feed device (18) is configured to move the reach mast (10) forwards and backwards transversely to the lifting direction (20), and wherein the camera (6) is mounted on or in the industrial truck (2) in such a way that the reach mast (10), as a vehicle component, is located at least partially in the image field of the camera (6), and wherein the evaluation unit (4) is configured to determine a pushing position of the reach mast (10) as a condition classification. [5] Industrial truck according to claim 4, wherein the evaluation unit (4) is further configured to determine, as a feature of the reach mast (10), a size of an image of at least a part of the reach mast (10) depicted at least in sections in the image data and to determine the pushing position of the reach mast (10) based on this feature. [6] Industrial truck (2) according to claim 4 or 5, wherein information on a geometry of the reach mast (10) and / or reference data on a plurality of push positions are stored in the evaluation unit (4), and the evaluation unit (4) is configured to determine the push position by evaluating the size of the image of at least a part of the reach mast (10) depicted in the image data, taking into account the information on the geometry of the reach mast (10), and / or to determine the push position by comparing the image information with the reference data, in particular taking additional account of the information on the geometry of the reach mast (10). [7] Industrial truck (2) according to claim 4 or 5, wherein the reach mast (10) has a machine-readable identification feature, in particular a QR code and / or an ArUco marker, as a feature, and the evaluation unit (4) is configured to determine a size of an image of the machine-readable identification feature in the image data, and to determine the pushing position of the reach mast (10) taking into account information on a geometry and / or size of the machine-readable identification feature, and / or to determine the pushing position of the reach mast (10) by comparing the image information with reference data, in particular taking into account the information on the geometry and / or size of the machine-readable identification feature. [8] Industrial truck (2) according to one of claims 4 to 7, wherein the evaluation unit (4) comprises a pre-trained neural network which is configured to receive the image data as input data and to output a pushing position of the reach mast (10). [9] Industrial truck (2) according to one of claims 4 to 8, wherein the evaluation unit (4) is configured to determine a plurality of push positions of the reach mast (10) for successive points in time and furthermore the evaluation unit (4) is configured to evaluate the plurality of push positions in a time-dependent manner with regard to a vibration of the reach mast (10), wherein the evaluation unit (4) is further configured to, when a vibration of the reach mast (10) is present, a) issue a warning message and / or b) to output control commands to a mast feed and / or travel control (22) which are suitable for actively counteracting the oscillation of the reach mast (10). [10] Logistics facility (24) with a control computer (26) and a plurality of industrial trucks (2) according to claim 3, which are designed to transmit the condition classification to the control computer (26). [11] Method for operating an industrial truck (2) with an evaluation unit (4) and at least one camera (6) which is mounted on or in the industrial truck (2) in such a way that at least one vehicle component of the industrial truck (2) is located at least partially in the image field of the camera (6), wherein the camera (6) is data-linked to the evaluation unit (4), comprising the following steps: Transferring image data from the camera (6) to the evaluation unit (4), Evaluating the image information contained in the image data in the evaluation unit (4) and determining features of the vehicle component and Carrying out the condition classification of the vehicle component or the camera (6) based on these features. [12] Method according to claim 11, wherein the evaluation unit (4) comprises a database comprising first features which occur in the image data when the vehicle component or the camera (6) is in a normal state, and second features which occur in the image data when the vehicle component or the camera (6) is in a deviating state, wherein the evaluation unit (4) carries out a state classification of the vehicle component or the camera (6) as “state: normal” or “state: deviating” based on a comparison of the image information with the first and second features or the evaluation unit (4) comprises a pre-trained neural network which receives the image data as input data and outputs a state classification of the vehicle component or the camera (6) as “state: normal” or “state: deviant”. [13] Method according to claim 12, wherein the evaluation unit (4) further outputs a message comprising the condition classification and transmits it in particular to a control computer (26) of a logistics facility (24) in which the industrial truck (2) is operated. [14] Method according to claim 11, in which a condition classification of the vehicle component is carried out on the basis of the features and the industrial truck (2) further comprises a reach mast (10), a load-handling means (8) and a mast feed device (18), wherein the load-handling means (8) is movable along the reach mast (10) in a lifting direction and the mast feed device (18) moves the reach mast (10) back and forth transversely to the lifting direction (20), and wherein the camera (6) is mounted on or in the industrial truck (2) in such a way that the reach mast (10) as a vehicle component is located at least partially in the image field of the camera (6), and wherein the evaluation unit (4) determines a pushing position of the reach mast (10) as a condition classification. [15] Method according to claim 14, wherein a size of an image of at least a part of the reach mast (10) depicted at least in sections in the image data is determined as a feature of the reach mast (10) and the push position of the reach mast (10) is determined on the basis of this feature. [16] Method according to claim 15, wherein information on a geometry of the reach mast (10) and / or reference data on a plurality of push positions are stored in the evaluation unit (4), and the evaluation unit (4) evaluates the size of the image of at least a part of the reach mast (10) depicted in the image data, taking into account the information on the geometry of the reach mast (10), and determines the push position, and / or by comparing the image information with the reference data, in particular taking into account the information on the geometry of the reach mast (10), and determines the push position therefrom or the evaluation unit (4) comprises a pre-trained neural network which receives image data as input data and outputs a push position of the push mast (10). [17] Method according to claim 14, wherein the reach mast (10) has a machine-readable identification feature, in particular a QR code and / or an ArUco marker, as a feature, and the evaluation unit (4) determines a size of an image of the machine-readable identification feature in the image data and, taking into account information on a geometry and / or size of the machine-readable identification feature, determines the pushing position of the reach mast (10) and / or determines the pushing position of the reach mast (10) by comparing the image information with reference data, in particular taking into account the information on the geometry and / or size of the machine-readable identification feature. [18] Method according to one of claims 15 to 17, wherein the evaluation unit (4) determines a plurality of thrust positions of the reach mast (10) for successive points in time and furthermore the evaluation unit (4) evaluates the plurality of thrust positions in a time-dependent manner with regard to an oscillation of the reach mast (10), wherein the evaluation unit (4) furthermore, when an oscillation of the reach mast is present, a) issues a warning message and / or b) outputs control commands to a mast feed and / or travel control (22) which actively counteract the oscillation of the reach mast (10).

Images