CARGO MONITORING PROCEDURES

Description

[0001] The invention relates to a method for detecting moving bodies or cargo parts inside a cargo space, wherein the cargo space is measured three-dimensionally with at least one distance-measuring, depth-sensitive sensor and the current distances of measuring points or sets of measuring points of any bodies or cargo parts located in the cargo space are detected by the sensor as measured values ​​in the form of 3D data.

[0002] The monitoring of cargo spaces using surveillance systems, especially electronic ones, is becoming increasingly important given today's traffic flows and the demands placed on freight logistics. The cargo spaces of trucks and containers are a particular focus for fleet operators, freight forwarders, and mail-order companies in this regard, and must be used as efficiently as possible and operated at a high level of security.

[0003] On the one hand, loading and unloading operations must be taken into account, which can be achieved by measuring, scanning or otherwise determining the bodies or parts of cargo that are loaded into a cargo space before or during loading.

[0004] On the other hand, with today's large-volume cargo spaces, it is essential to monitor the cargo for possible movements or changes in its position, especially for changes in position caused by transport operations, i.e., during the journey. This can be achieved through regular checks by the driver during the journey, but also through aids such as tensioning straps with tensile force measuring devices, or through automated processes for detecting changes in the position of cargo parts.

[0005] With regard to the first-mentioned detection of cargo items as such, methods are known in the prior art, for example, in which cameras are mounted on forklifts that scan the individual cargo items during loading and add the results to determine the total load or the fill level within the cargo space. Other solutions work with RFID tags / labels (RFID) attached to cargo items. radio frequency identification, Radio wave identification (RFID) is used, which is detected during loading or within the cargo space by corresponding antennas in the cargo space or, for example, in the loading area at the ramp. This makes it only possible to a limited extent to precisely determine a specific load condition within the cargo space.

[0006] In this regard, the prior art also shows methods in which individual cargo parts can be located and identified by size or dimensions using computer-aided measurements. EP 3 232 404 A1 discloses a method and a system for measuring the dimensions of a target object or a cargo part on a pallet or on a carrier. Using a sensor or a depth-sensitive camera, a 3D data set is determined that uses depth data, i.e., distance data (distance from the sensor), to describe the three-dimensional physical space occupied by the target object.

[0007] Computer-aided methods are also known in the state of the art for monitoring cargo for possible movements or changes in its position, which is important for safety reasons, i.e. with regard to the detection of moving bodies or parts of cargo inside a cargo space.

[0008] US 10,163,219 B2 discloses a method for displaying a changed load position, in which the position of a load part can be detected at an initial point in time with the aid of a camera and image processing software, for example by edge detection ( edge detection ) . The detected position is then displayed to the driver on a screen. The driver must then manually confirm the detection of the load and its position, for example by tapping the displayed position on the screen ( prompting ) .With the help of periodic further state detection runs using the camera and image processing software, it is then determined whether the position of the load part has changed and, if so, a signal is generated. Thus, a semi-automatic process can only detect that a change has occurred. However, a dynamic observation and interpretation of the type of movement is not possible.

[0009] US 2019 / 0065888 A1 discloses the monitoring of a cargo space using a depth-sensitive TOF camera (TOF: engl. time of flight ), with the help of which a cargo hold is first measured in depth, then the same cargo hold with additional or removed cargo parts, after which both images are finally compared and loaded or unloaded objects are determined. A more precise and reproducible description of the comparison procedure is not disclosed. The measurement always refers to a recording of the entire cargo hold scene, or to the total volume. No statement is made about the detailed behavior of a single cargo part.

[0010] The methods known in the prior art for monitoring movements or changes in the position of cargo parts are therefore only capable of detecting deviations in the position of a previously determined or manually identified cargo part and can generally only determine whether a movement or change in position has occurred, but not the dynamic course of this movement. This means that while the methods known in the prior art can detect that a movement / change has occurred, they cannot make a valid statement about the type of movement or change in position, for example, whether there is merely a swaying or wobbling of cargo parts, or whether the load is or has completely tilted or shifted / slipped.The detection is therefore reduced to the essentially binary statement of whether a movement has occurred or not.

[0011] Of course, the state of the art offers a range of generally applicable sensors and methods for detecting movement in a wide variety of application areas, for example in industry, domestic use, or security systems. A distinction is essentially made between active and passive motion sensors. Active sensors have a transmitter and a receiver and usually measure changes within the path of an emitted energy beam. Passive sensors react to changes in radiation in the observed environment, for example, changes in infrared radiation from people or objects. The use of such systems requires a relatively complex design and must be adapted to the objects or bodies to be monitored. If different objects are to be monitored, several sensors with different measuring principles may be required to make a reliable determination.The use of such motion detection systems, which are common in other areas, has not only not been adopted within cargo spaces for these reasons, but is also not the first choice from an economic point of view.

[0012] The object of the present invention is therefore to provide an improved method for monitoring cargo parts within cargo holds, which not only allows the position of bodies or cargo parts located in the cargo hold to be measured as such and their possible movements to be determined, but which also enables a dynamic analysis of the type of movement and with which the location of the bodies or cargo parts and the change in location within the cargo hold can be qualitatively determined. Furthermore, the object is to provide a method that can be applied to other surrounding systems without excessive intervention, which delivers sufficiently reliable results for determining and monitoring a cargo hold, and which can be used for any type of cargo hold.

[0013] This object is achieved by the features of the main claim. Further advantageous embodiments are disclosed in the subclaims. Also disclosed are a device for carrying out the method and a vehicle with a cargo space equipped with the device.

[0014] The cargo space is measured three-dimensionally using at least one distance-measuring, depth-sensitive sensor. In a first step, the 3D data describing each measurement point of any objects or cargo parts located in the cargo space are stored as a first data set, a so-called 3D point cloud, in a computing and / or evaluation unit and optionally provided with a timestamp. Such a timestamp is not absolutely necessary, since to represent the sequence of measurements, it is also possible to work with indexes / serial numbers assigned to each data set, with sequential storage / saving, or with fixed time intervals for the measurements.

[0015] In a second step, at least one further, subsequent three-dimensional measurement is carried out, whereby the 3D data describing the corresponding measuring points of the bodies or cargo parts located in the cargo space are stored in the computing and / or evaluation unit as additional data sets and optionally also provided with a time stamp.

[0016] Thereafter, either in a third step the 3D data of the individual, spatially corresponding measuring points of the first and the further measurement or measurements are compared with each other by means of an algorithm programmed in the computing and / or evaluation unit, wherein in the event that changes in the 3D data of spatially corresponding measuring points of the first and the further data set are detected, a fourth and further steps of the method are carried out as described below, or the fourth step described below is carried out without previously carrying out the third step.

[0017] An advantageous embodiment of the third step consists in comparing the 3D data from the first and subsequent measurements. If a number of changes in the 3D data exceeding a specified threshold is detected, the method steps explained below are executed. Considering such a threshold is useful, for example, to filter out random or unstable measurement data.

[0018] In the fourth step, the first and the further data sets (3D point clouds) are checked by means of an algorithm to determine whether subsets of the 3D data contained therein describe or characterize a three-dimensional body or a load part. If applicable, these 3D data are stored as a 3D data cluster (subset) characterizing a three-dimensional body or load part (volume), indexed and optionally also provided with a timestamp.

[0019] This is followed by a fifth step in which the 3D data clusters of the first and further data sets, which characterize three-dimensional bodies or charge parts determined from spatially corresponding measuring points or measuring point sets, are compared with one another using the algorithm in the computing and / or evaluation unit. If, during the comparison of the 3D data clusters for spatially corresponding three-dimensional bodies or charge parts of the first and further data sets, changes in the 3D data clusters are determined that characterize spatial changes in the spatially corresponding three-dimensional bodies or charge parts within a time interval, the computing and / or evaluation unit provides a signal describing these changes and capable of further processing.

[0020] The term "spatially corresponding measurement points or sets of measurement points" describes the measurement points within the boundaries of a space to be measured that are assumed to be at the same local position, or describe or occupy the same local position as in the previous measurement, thus essentially marking the same location in the three-dimensional coordinate system. To detect changes in the spatially corresponding measurement points, one can, for example, proceed by placing a three-dimensional expected space or range around the exact original determined value, within which a deviation may vary and, despite an absolute deviation in at least one dimension, is still considered a spatially corresponding, but "changed" measurement point.

[0021] Locally corresponding three-dimensional bodies or cargo parts are to be understood as volumes of any kind described by 3D data clusters and include a rectangular box as well as a body of any shape.

[0022] The comparison of the 3D data of the data sets in the third procedural step can also be performed in such a way that, instead of comparing the immediately preceding 3D data set with the subsequent data set, a more distant previous data set is used. This may be necessary if irregularities have occurred within a measurement cycle that suggest that the quality of the preceding data acquisition is inadequate.

[0023] An advantageous embodiment consists in that already during the first and second steps of the three-dimensional measurement of the bodies or cargo parts located in the cargo hold, their parameters, in particular with regard to their position and their respective width, height and depth, are stored and indexed as 3D data clusters. Within the scope of such an advantageous procedure, a certain variation in the sequence of the processing steps is possible, whereby, for example, as described above, the fourth method step is also made possible independently of the third method step. A variation may essentially comprise a summary or combination of method steps and, in this sense, does not depart from the original scope of the method according to the invention, i.e., does not constitute an aliud.

[0024] Coming back to the first step, the interior of the cargo space is scanned with a distance-measuring, depth-sensitive sensor ( depth-aware sensor ) are measured three-dimensionally. For this purpose, the 3D data describing each measurement point of any bodies or cargo parts located in the cargo space are stored as a first data set (3D point cloud) and optionally provided with a timestamp. In the second step, at least one further, subsequent, corresponding three-dimensional measurement is then carried out using the sensor. In this subsequent measurement, the 3D data describing the corresponding measurement points are also stored as an associated second or further data set and optionally also provided with a timestamp. The at least one further three-dimensional measurement is to be understood here as "each further" measurement in the sense of each repeated subsequent measurement (second or further data set) to a previous measurement (first data set).

[0025] The 3D data also represents so-called depth data ( depth data ), namely the "depths," or more precisely, the distances of the corresponding measurement points from the sensor. This allows one to determine the distances of each detected surface of an object and thus its position. Repeating the measurement allows one to determine the change in this data.

[0026] Based on the consideration that, through measurement using a depth-sensitive sensor, not only the distance of each measuring point belonging to a body or piece of cargo to the sensor is known, but also the distances between individual measuring points. Based on the assumption that a small spatial change in a measuring point also generates a change, albeit minimal, in the associated 3D datum when comparing two consecutive data sets, the idea underlying the inventive method is that an initially easily detectable "individual" change triggers further, more comprehensive processing or method steps when individual changes become noticeable. Such "individual" changes are detected very quickly and precisely in either the third or fourth method step.

[0027] The change in individual 3D data is thus assigned a kind of indicator function for potentially more serious changes to be expected. This is particularly successful when a change in individual, particularly conspicuous and prominent 3D data is detectable, for example, those belonging to the border or corner areas of a cargo part.

[0028] If such changes occur, further processing steps are then performed to determine whether the individual 3D datum or the number of modified 3D data items belongs to 3D data that describe a cargo part or a body (a volume). Using indexing, optionally including a timestamp, it is further determined whether, in which direction, and at what speed the cargo part or the body (the volume) to which this individual 3D datum or the number of modified 3D data items belongs has moved.

[0029] The inventive method leads to very fast and reliable detection of changes in the cargo parts or volumes. In an advantageous further development, the depth, height, and width of the moving cargo parts or volumes are also determined and stored as a 3D data cluster, i.e., as a subset of the respective data set. Such a three-dimensional evaluation method makes the old and new positions of the cargo part clearly recognizable, as well as the direction and type of movement.

[0030] The measurements with the distance-measuring, depth-sensitive sensor can be carried out periodically, which is advantageous with regard to continuous status determination, but they can also be carried out at irregular intervals or as required.

[0031] A further advantageous embodiment consists in that in the third step the 3D data is acquired in such a way that, on the one hand, 2D data of a 2D image acquired by a sensor or a camera is determined and, on the other hand, distance information / a distance is determined for each pixel of the 2D image. In the third step, only the 2D data of the individual, spatially corresponding measuring points of the first and the further measurement or measurements are compared with one another using an algorithm programmed in the computing and / or evaluation unit. If changes in the 2D data of spatially corresponding measuring points of the first and the further data set are detected, in particular if a number of changes exceeding a predetermined threshold is detected, the further method steps are carried out by incorporating the distance information or the 3D data for each measuring point.

[0032] Such an evaluation, initially reduced to 2D data only, represents a simplified initial test, which can be carried out using simplified software. This test can determine whether changes in the 2D data are already present in a two-dimensional view. If so, the subsequent procedural steps are then carried out using the 3D data.

[0033] In this embodiment, in the third step, the pixels of a two-dimensional representation of the individual, spatially corresponding measuring points of the first and the further measurement or measurements are compared with each other by means of an algorithm programmed in the computing and / or evaluation unit and, in the event that a number of changes in the pixels of spatially corresponding measuring points of the first and the further data set or one of the further data sets is detected that exceeds a predetermined threshold value, the further steps are carried out.

[0034] This also reduces the amount of data to be processed for the first detection step according to the method according to the invention, as only pixels, i.e., image points of a two-dimensional representation of the measured values, are compared with each other. This also provides an indicator function that triggers the subsequent process steps for the precise determination of the movement.

[0035] A further advantageous embodiment consists in checking the change in the 3D data compared to the previous measurements for a change in the positioning of a body or load component in the spatial x-, y-, or z-direction with the same or changing sign. Particularly in conjunction with a further advantageous embodiment, in which the change in the 3D data compared to the previous measurements is checked for a spatial change in a 3D data cluster characterizing an indicated three-dimensional body or load component, a highly dynamic and rich interpretation of the obtained data is achieved.

[0036] With such direction determination and identification by recognizing the cluster indexing, the method according to the invention is capable of determining whether any swaying or wobbling of cargo parts is noticeable and, in a further comparison of the associated indexed data clusters, whether this swaying has led to a cargo part completely changing its position, for example, falling from a stack. The method according to the invention is thus capable of detecting not only swaying or wobbling of the cargo, but also a subsequent change in the position of a body or cargo part, such as falling or sliding within the cargo space. All of this is possible in almost real time using the method according to the invention.

[0037] This makes it possible to determine very precisely and in real time whether a previously identified cargo item or a previously identified object has moved, for example, slipped. Furthermore, even very small movements of the cargo can be detected, allowing a quick assessment of the cargo's security status. If a cargo item falls from a stack of high-stacked goods, the method according to the invention can also determine the height from which the cargo item fell and its current position after the fall.

[0038] The further processable signal can be used to generate warnings for the driver, to generate signals for a further processing control unit or to generate data for a visual representation of the loading space, the spatial allocation of the bodies or load parts in the loading space and the current loading state, in particular for display on a monitor or a display device, in particular as a three-dimensional load image.

[0039] The provided signal, which can be further processed, can also contain all parameters calculated from the 3D data, such as those that describe the position and dimensions of bodies or parts of the load.

[0040] The 3D data can be used as a basis for further model calculations or for generating video data, which can then be displayed on a monitor and produce a three-dimensional image depicting the current cargo status. The data or its processing can be sent via communication devices, for example, to the headquarters of a freight forwarding company or to authorities responsible for inspecting the cargo.

[0041] To create a direct relationship to an observed cargo space, the algorithm can be designed so that the 3D data of the cargo space's boundary surfaces, i.e., the floor, side walls, and ceiling, are also stored in identifiable 3D data clusters and are thus known and "set" for all further measurements. This information is then used to easily distinguish between 3D data describing the cargo space as such and 3D data describing the cargo parts or bodies.

[0042] Advantageously, the zero point of the coordinate system is set in a corner of the cargo space opposite the sensor. The sensor is advantageously positioned on a wall or boundary of a cargo space and directed toward the opposite wall so that the sensor's measuring range encompasses the entire cargo space, with the zero point of the three-dimensional measurements being calibrated to a corner of the cargo space. Any other point can also be used as the zero point of the coordinate system. With such a sensor arrangement, the entire cargo space up to the front wall is within the measuring range.

[0043] The method according to the invention can also be used to determine whether an object or cargo part has been added to the cargo hold, removed, or moved from its original location to another location within the cargo hold. This simply requires comparing the 3D data sets and 3D data clusters of a 3D data set determined before the event (adding a body or cargo part) with a data set containing 3D data measured after the event. New objects or cargo parts added within the cargo hold are then assigned their corresponding 3D data clusters, which can be identified through further measurements.

[0044] Another advantageous design is that the sensor is configured as an optical depth sensor, particularly as a time-of-flight (TOF) camera or stereo camera. A TOF camera, for example, provides the distance of the body depicted on it for each pixel. This allows an entire scene to be recorded at once without having to scan each individual scene. This naturally leads to faster processing of the corresponding signals. Depending on the application, it is also advantageous if the sensor is configured as a LiDAR sensor ( light imaging, detection and ranging ) or laser scanner ( light amplification by stimulated emission of radiation ), which represents a more cost-effective alternative for raster scanning. These systems are well known.

[0045] A further advantageous embodiment consists in checking the confidence level and completeness of the modified 3D data of the second data set and / or setting a threshold for noise in the measurement.

[0046] This can be achieved, for example, by evaluating the temporal change of the measured values ​​characterizing the distances between measuring points or sets of measuring points. From the temporal history, it can be determined based on threshold values ​​or occurrences / occurrences whether, for example, an expected measured value—i.e., the occurrence of a measuring point described by 3D data in a certain expected range, i.e., at a certain distance—only occurs after a time threshold has been exceeded, or whether it only appears for a short period of time and then disappears again, i.e., does not remain stable.

[0047] Through the described three-dimensional measurement, the currently determined state and the previous state determined during the previous measurement are known and thus also the parameters of the corresponding bodies or parts of the load, namely their depth (distance from the sensor), their width, their height, their position and their volume.

[0048] By comparing the currently acquired 3D data from a 3D data cluster with the existing 3D data cluster information and performing additional checks to determine whether the 3D data has a minimum quality or stability, or whether it does not appear frequently enough within an expected range, it is also possible to determine whether the sensor was unable to measure certain areas. Similarly, the 3D data surrounding the respective measurement points can be analyzed and compared with the previous data.

[0049] A further advantageous feature is that a visual representation of the cargo space, the location of the objects or cargo parts within the cargo space, and the current loading status are displayed on a monitor in the form of a three-dimensional cargo image. A corresponding computer processing of the 3D data then produces easy-to-interpret images on a monitor, allowing a very quick and intuitive assessment by a person inspecting the cargo.

[0050] A further advantageous embodiment consists in the periodic measurement of the 3D data of each measuring point being carried out at a frequency of 1 Hz to 5 Hz. Such a periodic check is sufficiently dynamic to achieve sufficient accuracy for monitoring a truck loading space and a defined determination of movements of a load part when applying the method according to the invention.

[0051] It makes sense to use the 3D data, which describes the parameters - i.e. the width, height and depth - of each individual body or piece of cargo in the cargo space to determine the space required and / or the volume of the bodies or pieces of cargo and to relate this to the available cargo space area or to the available volume of the cargo space. The total occupied volume can then be calculated, for example, from the sum of all calculated volumes of all bodies and / or pieces of cargo and thus also the remaining free volume in the cargo space. The same applies to calculating the total area occupied by all bodies and / or pieces of cargo in relation to the remaining free, unoccupied area. The relevance of such specifications for load planning by fleet operators or freight forwarders, for example, is immediately apparent.

[0052] A further advantageous embodiment consists in providing the further processable signal for storage in data processing systems, particularly for use in control systems and for use and processing within a data communication system. Thus, these signals can also be sent to the headquarters of a freight forwarding company or fleet operator via any additional radio devices connected to the vehicle belonging to the cargo area.

[0053] Furthermore, a device is disclosed which is suitable for monitoring a cargo space and for carrying out the method according to the invention. The device comprises at least one distance-measuring, depth-sensitive sensor for repeated three-dimensional measurement of the cargo space, as well as a computing and / or evaluation unit with a programmed algorithm. By means of the computing and / or evaluation unit thus provided and the algorithm, the 3D data of each measuring point of any bodies or cargo parts located in the cargo space are determined through repeated measurements, stored as data sets (3D point clouds), and compared with one another according to the method according to the invention. If changes are detected, a processable signal is provided.The advantage lies in the usability / provision of such a facility for any loading space, for example for the loading spaces of trucks, but also for loading containers of ships or other vehicles.

[0054] The method according to the invention and the device particularly suited thereto are particularly suitable for a vehicle with a cargo space, for example, a truck or a train of vehicles. Application to containers transported in different vehicles is also very possible. The respective vehicle or container has a cargo space equipped with at least one distance-measuring, depth-sensitive sensor for three-dimensional measurement located inside the cargo space. The vehicle further comprises a computing and / or evaluation unit with a programmed algorithm with which the 3D data of each measuring point acquired by the sensor is stored in a first data set (3D point cloud), processed according to the method according to the invention, and compared with at least one further data set.

[0055] The invention will be explained in more detail using an exemplary embodiment. Fig. 1 shows a sketch of a monitor representation of a cargo space based on recorded 3D data which was processed computationally to display a three-dimensional cargo image, Fig. 2 shows a representation according to Fig.1 , however, there a situation in which a load part is tilted and shifted from its original position, Fig. 3, 4 a two-dimensional, simple monitor representation of the result of the third step of the method according to the invention, Fig. 5, 6 an enlarged view of the in Fig. 3 and 4shown details in their overall environment, Fig. 7 the result of the further inventive method steps prepared for a monitor display, in which the first and the further data set or the further data sets (3D point cloud) are checked by means of an algorithm to determine whether subsets of the 3D data contained therein describe or characterize a three-dimensional body or a cargo part.

[0056] Fig. 1 shows a sketch of a monitor display in which, with the aid of the method according to the invention, a cargo space was measured three-dimensionally with at least one distance-measuring, depth-sensitive sensor. The current distances from measuring points or measuring point sets of any objects or cargo parts located in the cargo space were recorded by the sensor as measured values ​​in the form of 3D data and processed mathematically for the most realistic representation of the loading condition in the form of a three-dimensional cargo image on a monitor. Such processing of the 3D data then provides the easily interpretable image according to Fig. 1 on a monitor.

[0057] A distance-measuring TOF sensor, not shown in detail here, is located in an upper corner of the rear tail lift of cargo space 1. This location is particularly well suited for the installation of such a sensor, since the entire cargo space up to the front wall is then in the measuring range of the sensor and the measuring range is not obstructed by any objects or cargo parts that are in front of it, as long as the cargo is always stowed properly down to the depth of the cargo space.

[0058] Fig. 1 shows a view of the interior of a loading space 1, namely from its rear tail lift towards its front boundary wall. The right boundary wall 3, the left boundary wall 4, the loading area or floor 5, and the front boundary wall 2 can be seen. Various objects are also visible, namely various cargo items 6 located on the loading area floor 5. This is the loading space of a truck, the loading area of ​​which is provided with a framework of struts and slats, which is covered with a tarpaulin.

[0059] Fig.1 shows an initial situation in which the cargo part 6.1 is in its original position, namely in the position in which it was loaded.

[0060] In contrast, Fig. 2 a situation in which the load part 6.1 is tilted and shifted from its original position. Otherwise, the representation and data processing in Fig.2 those in Fig.1 .

[0061] Fig. 3 now shows in a different and merely two-dimensional simpler monitor display the result of the third step of the method according to the invention, in which pixels, namely 2D data of the individual, spatially corresponding measuring points of the first and the further measurements were compared with each other by means of an algorithm programmed in the computing and / or evaluation unit. Fig. 3 and 4 each represent enlarged sections of the overall picture of the Fig.1 and 2 in a view tilted to the left.

[0062] Looking at the representations of the Fig.4 compared to the representation in Fig.3 , you can see that in Fig.4 a number of changes in the pixels or 2D data of spatially corresponding measuring points of the first and the further data set were detected, namely a change in individual, particularly conspicuous and prominent 2D data 7 belonging to the boundary or corner area of ​​the load part 6.1.

[0063] Fig. 5 and 6 show a slightly enlarged view of the Fig. 3 and 4 shown details in their overall environment, whereby here also a left-tilted view of the cargo space is present. In the Fig. 5 and 6 Auxiliary lines calculated from the other determined data are also inserted, which make the loading space walls 2, 3, 4 and the loading space floor 5 visible.

[0064] Fig. 7Finally, the result is displayed on a monitor, after which the further inventive method steps are triggered. The detection of changes in individual data then leads to the first and subsequent data sets (3D point cloud) being checked by an algorithm in the fourth method step to determine whether subsets of the 3D data contained therein describe or characterize a three-dimensional body or a cargo part.

[0065] This applies here to load part 6.1, so that the 3D data of load part 6.1 was stored as a 3D data cluster (subset) characterizing this three-dimensional load part, indexed, and also provided with a timestamp. In the further procedural step, namely the fifth step, the 3D data clusters of the successive, different measurements are then compared with each other, after which spatial changes of the load part within a time interval were detected, and a signal describing these changes and suitable for further processing was provided by the computing and / or evaluation unit. In this case, the signal is a warning signal to the effect that, after an initial fluctuation of the load or load part 6.1, it was ultimately determined that the load part is falling or has fallen from its stack. List of reference symbols (part of the description)

[0066] 1Cargo space 2Front wall of the cargo space 3Right boundary wall of the cargo space 4Left boundary wall of the cargo space 5Cargo space floor 6Cargo items, entire cargo 6.1Individual cargo item 7Change in individual conspicuous 3D data

Claims

1. Computer-implemented method for detecting moving bodies or pieces of cargo inside a cargo space, wherein the cargo space (1) is measured three-dimensionally using at least one distance-measuring, depth-sensitive sensor, and the current distances of measuring points or measuring point sets of any bodies or pieces of cargo (6, 6.1) located in the cargo space are detected by the sensor as measured values in the form of 3D data, - wherein, in a first step, in a first measurement, the 3D data describing each measuring point of any bodies or pieces of cargo (6, 6.1) located in the cargo space are stored as a first data set (3D point cloud) in a computing and / or evaluation unit and optionally provided with a time stamp, - wherein, in a second step, at least one further, subsequent three-dimensional measurement is executed, and the 3D data describing the corresponding measuring points of the bodies or pieces of cargo (6, 6.1) located in the cargo space are stored in the computing and / or evaluation unit as further data sets and optionally also provided with a time stamp, - wherein, in a third step, 3D data of the individual, spatially corresponding measuring points of the first and the further measurement or measurements are compared with one another by means of an algorithm programmed in the computing and / or evaluation unit, and, in the event that changes in the 3D data of spatially corresponding measuring points of the first and the further data set are detected, the fourth step and the following further steps are executed, characterized in that - without prior execution of the third step, in a fourth step, the first and the further data set / the further data sets (3D point cloud / 3D point clouds) are checked by means of an algorithm programmed in the computing and / or evaluation unit to ascertain whether subsets of the 3D data contained therein characterize a three-dimensional body or a piece of cargo (6.1) and, if applicable, are stored as a 3D data cluster characterizing a three-dimensional body or piece of cargo (volume), indexed and optionally also provided with a time stamp, - wherein, in a fifth step, the 3D data clusters of the first and further data sets, which data clusters characterize three-dimensional bodies or pieces of cargo determined from spatially corresponding measuring points or measuring point sets, are compared with one another by means of the algorithm in the computing and / or evaluation unit, wherein, in the event that, during the comparison of the 3D data clusters for spatially corresponding three-dimensional bodies or pieces of cargo of the first and further data sets / further data sets, changes in the 3D data clusters are determined that characterize spatial changes in the spatially corresponding three-dimensional bodies or pieces of cargo within a time interval, a signal which describes these changes and can be further processed is provided by the computing and / or evaluation unit.

2. Method according to claim 1, in which, in the third step, the 3D data of the first and the further measurement / further measurements are compared with one another, and in the event that a number of changes (7) in the 3D data exceeding a predetermined threshold value is detected, the steps according to claim 1 following the third step are executed.

3. Method according to claim 1 or claim 2, in which the interior of the cargo space (1) is periodically measured three-dimensionally.

4. Method according to any of claims 1 to 3, in which, during the first and second steps of the three-dimensional measurement of the bodies or pieces of cargo located in the cargo space, the parameters of said bodies or pieces of cargo, in particular with regard to their position and their width, height and depth, are stored as 3D data clusters and optionally indexed and / or provided with a time stamp.

5. Method according to any of claims 1 to 4, in which the 3D data are detected in such a way that 2D data of a 2D image detected by a sensor or a camera are determined and that distance information is determined for each image point (pixel) of the 2D image, wherein, in the third step, only the 2D data of the individual, spatially corresponding measuring points of the first and the further measurement / further measurements are compared with one another by means of an algorithm programmed in the computing and / or evaluation unit, and, in the event that changes in the 2D data of spatially corresponding measuring points of the first and the further data set / further data sets are detected, the further method steps are executed by incorporating the distance information or the 3D data for each measuring point.

6. Method according to any of claims 1 to 5, in which the change in the 3D data compared to the previous measurements is checked for a positioning of a body or piece of cargo that changes with the same or varying signs in the spatial x-, y-, or z-direction.

7. Method according to any of claims 1 to 6, in which the change in the 3D data compared to the previous measurements is checked for a spatial change in a 3D data cluster characterizing an indexed three-dimensional body or piece of cargo.

8. Method according to any of claims 1 to 7, in which the sensor is designed as an optical depth sensor, in particular as a time-of-flight camera or stereo camera.

9. Method according to any of claims 1 to 8, in which the sensor is designed as a LiDAR sensor or laser scanner.

10. Method according to any of claims 1 to 9 for monitoring a cargo space belonging to a vehicle, in which the sensor is arranged on a wall or boundary of a cargo space, and the measuring range of the sensor comprises the entire cargo space.

11. Method according to any of claims 1 to 10, in which the confidence level and completeness of the changed 3D data of the further data set or of one of the further data sets is checked, and / or a threshold value for noise during the measurement or classification is set.

12. Method according to any of claims 1 to 11, in which a visual display of the cargo space, the spatial allocation of the bodies or pieces of cargo in the cargo space and the current cargo status is displayed on a monitor in the form of a three-dimensional cargo image.

13. Method according to any of claims 2 to 12, in which the periodic measurement of the 3D data of each measuring point is executed at a frequency of 1 to 5 Hz.

14. Method according to any of claims 1 to 13, in which the further processable signal is provided for storage in data processing systems, in particular for use in control devices and for use and processing within a data communication system.

15. Device for monitoring a cargo space, for carrying out the method according to claims 1 to 14, wherein the device comprises at least one distance-measuring, depth-sensitive sensor for repeated, three-dimensional measurement of the cargo space as well as a computing and / or evaluation unit which has a programmed algorithm, wherein by means of the computing and / or evaluation unit and the algorithm, the 3D data of each data set obtained by repeated measurement, which 3D data describe each measuring point of any bodies or pieces of cargo (6, 7, 8, 11) located in the cargo space, are compared with one another using a method according to claims 1 to 14, and, in the event of detected changes, a processable signal is provided.

16. Vehicle comprising a cargo space, in particular a truck or a train convoy, wherein the vehicle comprises a device for monitoring a cargo space according to claim 15.