EP2888185A1

EP2888185A1 - Method and system for assisting a worker in a system for manipulating goods items

Info

Publication number: EP2888185A1
Application number: EP13780049.6A
Authority: EP
Inventors: Markus Winkler
Original assignee: TGW Logistics Group GmbH
Current assignee: TGW Logistics Group GmbH
Priority date: 2012-08-24
Filing date: 2013-08-23
Publication date: 2015-07-01
Also published as: AT513130A4; WO2014028959A1; AT513130B1

Abstract

A method and an assistance system for assisting a worker (1) in a system (2) for manipulating goods items (3) are specified. According to the invention, the movements of the worker (1) and optionally the goods item (3) which is manipulated by the worker (1) are detected by way of at least one camera (9) which is mounted on the system side, and are compared as to whether the detected movements of the worker (1) and optionally of the goods item (3) lie in a target range of a movement sequence. If this is the case, the performance of a further movement sequence is subsequently checked. If this is not the case, a correction signal is output in order to bring about the performance of the required movement sequence. To this end, the detection region of the at least one camera (9) is divided into segments, wherein a movement sequence contains an occupancy of the segments by the worker (1). In general, a movement sequence comprises the detection and depositing of at least one goods item (3).

Description

Method and system for assisting a worker in a facility for manipulating goods

The invention relates to a method and an assistance system for assisting a worker in a facility for manipulating goods. In addition, the invention relates to a use of said method and / or said assistance system.

Methods and assistance systems of the type mentioned are known in principle. For example, EP 1 487 616 B1 discloses an augmented reality system for a warehouse. Here, actual movements of a warehouse worker are compared using a camera with target movements and logged accordingly. The necessary camera is mounted on the helmet of a worker. Optionally, additional information, such as the location of a part to be detected, may be displayed in data glasses.

EP 2 161 219 B1 discloses another augmented reality system for a warehouse worker where the camera is also mounted on a worker's helmet. In particular, a kind of navigation system can be realized with a data goggles, in which a target movement of the worker is displayed by means of superimposed directional arrows.

Finally, WO 03/034397 A1 discloses an augmented reality system which can be used inter alia in a warehouse. The camera is in turn mounted on the helmet of a worker. Although the presented systems make work easier for a warehouse worker, these systems have narrow limits. For example, the movements of a worker can not be monitored independently of his line of sight. The monitoring of a workflow is dependent on the worker and his helmet camera looking at his hands. However, this can not be assumed in any case, because in particular experienced workers do things "blindly", without necessarily looking at their hands, and workers are frequently observed who perform their work without any problems, despite lively conversations with their colleagues. len. Her gaze often wanders to the other party and is rarely focused on the actual job.

Another disadvantage of helmet cameras is their weight, which sometimes seriously affects the productivity of the worker. Especially with aggravating working conditions such as high temperature and high humidity wearing such a helmet is perceived as very uncomfortable. In any case, this increases the sweat on the head. If the worker wipes sweat from his forehead, the camera is "blinded" for a few moments and can not continuously monitor the workflow, meaning important steps can be "overlooked" by the camera. When the work is done in a cold environment, the worker's breathing air often causes the camera lens to fog up, which also makes monitoring of a workflow impossible. Finally, the weight of the camera and data glasses limits the maximum head acceleration that occurs, for example, when turning the head fast. Therefore, the worker can not work as fast as his bodily physiology may allow.

Determining the position of the worker with the aid of a helmet camera is difficult, in particular because of the sometimes violent movements and the "blind" moments of the camera If the system "out of gear", that is loses its position, it takes a certain time until the Position and orientation of the helmet camera is detected, and the system again can perform a meaningful monitoring of a workflow. All in all, depending on the computing power and behavior of the worker, such a system can slow down rather than accelerate the work to be done. The monitoring of various activities occurring in a warehouse is hardly possible to impossible with the known devices, for example the manipulation of objects which are difficult to distinguish. As a striking example, the re-sorting of screws from one red container into another red container, which also contains the same screws, is mentioned. As can be seen from the limitations already mentioned, the meaningful monitoring of such an operation is practically impossible. Assuming that the camera loses sight of the containers even for a single moment, the assignment of the source and target containers when the view is restored and thus the monitoring of the work process is extremely difficult with the utmost security. It is not possible to determine if the worker did not swap the source and target containers during a "blind" phase of the camera to improve his / her working situation, the camera's image area covers only one of the containers (such as when using a relatively long focal length lens in relation to the size of the camera) Container), then the unambiguous determination, which container is just detected, is also virtually impossible.

The known systems are primarily set up for the operations to be performed by a worker. If several people work together, meaningful monitoring of the workflow is very difficult and only possible with very high computational effort. If, for example, three people work together, the work area is covered by three different cameras whose coverage is influenced by the head movements of the persons concerned. The ability of people to look a conversation partner in the eye is also a major hindrance to monitoring such a workflow. If one of the three persons gives instructions to the other two, they (and thus their helmet cameras) inevitably look at the instructor. Although, in principle, there are three cameras for observation in this work situation, none of them detects the object that is actually manipulated (e.g., shared or handed over).

Restrictions of a worker, which are given due to his physique, can not be considered with the known systems. Therefore, it may happen, for example, that a worker is instructed to remove goods from a shelf that is actually too high for him. As a result, the worker in question must laboriously obtain an ascent aid or ask a colleague to do the work for him. Both are inefficient, but the latter also leads with high probability to disturbances in the monitoring of the work process, since a person other than the one provided by the system does the work.

The object of the present invention is now to specify an improved method and an improved assistance system for assisting a worker in a facility for manipulating goods. In particular, the above-mentioned disadvantages are to be overcome.

The object is achieved by a method for assisting a worker in a device for manipulating goods, comprising the steps: a) detecting movements of the worker with at least one camera mounted on the installation side,

b) comparing whether the detected movements of the worker are within a desired range of a first movement sequence,

c) outputting a correction signal and continuing at step a) with the first motion sequence if the comparison is negative or continuing at step a) with a second motion sequence if the comparison is positive, wherein a motion sequence comprises detecting and discarding at least one commodity, the detection range of the at least one camera is subdivided into segments, and wherein a motion sequence contains occupancy of the segments by the worker.

The invention is further solved by an assistance system for assisting a worker in a device for manipulating goods, comprising:

at least one camera mounted on the system for detecting movements of the worker,

Means for comparing whether the detected movements of the worker are in a desired range of a movement sequence, and for outputting a correction signal if this is not the case, wherein a movement sequence comprises detecting and storing at least one product, wherein the detection range of at least one camera in Segments is divided and wherein a motion sequence contains occupancy of the segments by the worker.

By detecting both the movement of a worker from a remote observer site, the system can be used very flexibly under changing conditions and detects the monitored workflow with high security. This significantly improves the monitoring of a workflow. In particular, the following advantages result in particular:

The movements of a worker can be monitored independently of his line of sight. Thus, the system can also be used for "experienced" workers who have not kept their eyes on their hands.

The worker is more efficient because he is not affected by the weight of a helmet camera and can move freely. For example, fast head turns are also unproblematic for the presented system in terms of image processing. That too Fogging of the camera by the worker's breathing air is unlikely because the camera may not necessarily be mounted directly above the worker's head but at a more advantageous location, such as slightly away from him.

The position of the observing camera is always known and does not have to be laboriously calculated. Therefore, the system can not be "out of step." A specification of work steps can therefore be made fluid even with comparatively low computational effort.

Monitoring the manipulation of hard-to-distinguish items is facilitated. For example, re-sorting screws from one red container into another red container, which also contains the same screws, can always be detected correctly. This is the case even if the worker exchanges source and destination containers, for example. In particular, the said method or the said assistance system can therefore be used to monitor a sequence of movements which relates to the picking and / or resorting of goods.

- It is easier to monitor several workers who work together and, for example, hand over objects or carry an object together, because the observing camera can always capture the entire scene.

Limitations or benefits of a worker due to his physique may be taken into account, as features such as height, arm length, and stature of the worker may be captured by the camera. As a result, work can be assigned to the most appropriate person, for example working on high shelves to tall people and manipulating heavy objects on people with strong physique. Of course, work can also be assigned gender-specific. Advantageously, the said method or the said assistance system are used in a warehouse. A warehouse is used for receiving, storing and usually distribution of goods. A warehouse can be designed, for example, as a high-bay warehouse, small parts warehouse, floor storage, high-level storage or as a paternoster storage and, for example, shelves of various types (such as shelf shelves, pallet racking, continuous shelves and high shelves) as well as various types of funding (such as roller conveyors, belt conveyors, forklifts, and driverless transport vehicles). Due to the segmentation of the detection range of the at least one camera, the calculation effort for the said method is advantageously kept low. For example, a grid can be laid over a work surface (worktable) and preferably also displayed by a laser projector. Of course, fixed markings on this work surface are also conceivable. Movements are evaluated as a sequence of occupied segments.

In the context of the invention, a "camera" is understood to mean a device or system which images a captured scene two-dimensionally or three-dimensionally.A "camera" can capture a scene by detecting the light generated and / or reflected by the scene - both in the scene visible as well as in the invisible wavelength range - image. However, a "camera" within the meaning of the invention does not necessarily use light as an information carrier, but additionally or alternatively, the detection of a scene can also take place in other ways. Image of a scenery on

Based on the duration of emitted and reflected by the scenery rays or waves generated. For example, this can be done with the help of infrared rays, ultrasound and / or laser beams. It is also conceivable that a plurality of cameras are provided, each of which captures only a subarea of the scene and thus a subarea and / or a subset of the segments. For example, one camera may capture a subset of the segments, another camera may capture a different subset of the segments. The subsets may overlap or not. It is also conceivable that one camera ever captures a segment. But it would also be possible for (large) segments to be detected jointly by several cameras. Of course, mixed forms are conceivable.

Further advantageous embodiments and modifications of the invention will become apparent from the dependent claims and from the description in conjunction with the figures.

It is advantageous if the movements of the goods, which is manipulated by the worker, are detected with the at least one camera before step b) and it is compared in step b) whether the detected movements of the worker and of the goods are in a setpoint. Range of a first movement sequence are. The fact that not only the worker is detected by the camera, but also the goods manipulated by him, the monitoring of movement is particularly safe. In this embodiment, it is not sufficient, for example, to move only the hand in a predetermined manner. Instead, the goods must be moved in a predetermined manner.

It should be noted at this point that the detection of the goods by the camera can of course not be ruled out even if it is merely compared whether the detected movements of the worker lie within a desired range of a first movement sequence. In this case, however, the goods may be detected "incidentally" and without further consequences, whereas in the case of the preceding embodiment the goods are recorded as planned and their movement is evaluated.

It is particularly advantageous if a product is classified as being "detected" by the worker, if it is detected between the detection and depositing, in particular during the entire sequence of movements, at least partially by a hand of the worker or a tool held by him (eg pliers, gripper, As a rule, a commodity is at least partially obscured when it is held by a worker, and this circumstance is exploited in the present embodiment to classify a commodity as "captured". If it is not completely recognized between the detection (start point of the movement sequence) and the departure (end point of the movement sequence), then it can be assumed with high probability that it will be moved by the worker in the given way and not, for example, only between start and end point lying on a work table. It is checked whether a product during the entire movement sequence at least partially by a hand of the worker or a held by him

If the tool is hidden then it can also be avoided that a product is classified as being grasped by the worker, for example if a worker's hand is merely moved over a product lying on the work table. To determine whether a product is at least partially obscured, for example, data from a database in which 3D models of the goods are stored can be used. If a view of the relevant 3D model does not completely correspond to the view of the goods captured by the camera, and in particular corresponds to the hidden portion (at least in part) of a worker's hand or a tool held by him, then the goods can be classified as recorded.

It should be noted at this point that the aforementioned variant of the method (recognition of at least partially concealed goods) can form the basis for an invention independent of the features of patent claim 1. In particular, in these cases, a segmentation of the detection range of the camera can be dispensed with. Analogously, this also applies to a corresponding device or the use of such a method and / or a corresponding device in a warehouse, in particular when picking and / or resorting of goods.

It is advantageous if, prior to step a), at least the starting point and / or the end point of a movement sequence are displayed. In this way, the worker is shown what to do. For example, the starting point and the end point may be provided with different visual markings to indicate to the worker what to do. For example, the start and end points can be marked with different colors. It is particularly advantageous if the camera identifies persons and performs the display of at least the start and end points of a movement sequence on the basis of a personal setting (profile). Here, preferences but also restrictions of certain users can be taken into account. For example, by default, the starting point and the end point may be marked red and green, but purple and orange, for example, if the worker has poor eyesight in distinguishing red and green. For color-blind people, the starting point and end point can also be marked with different shapes, such as a circle and a cross. In general, for example, indicator lamps or laser projectors are available for the optical marking of the start and end points, which can be used in a particularly flexible manner for producing different markings by means of the beam deflection. In addition, start and end point can also be done with the help of a screen or a corresponding insertion in a data glasses. Especially with screens and data glasses, the display of even complex movement sequences is possible. For example, a worker may be instructed to rotate a detected item before it is discarded. In addition to visual markings or in addition, of course, acoustic instructions are also possible, for example "T-shirt from the left stack, turn 90 ° to the left and in shipping box number 1 It is also possible to mark the starting point and / or end point of a sequence of movements with a number and / or type (if a container contains different types of goods) of the goods to be detected / stored, for example with corresponding LED displays, LCD Finally, it is also conceivable that the start point and end point of a movement sequence are not displayed at the same time, but with a time offset.The time offset can be fixed (eg one second) or individually calculated / estimated on the basis of the motion sequence It is also possible to show only the starting point for the time being and the end point only when it is recognized that the goods have been taken from the starting point.

It is advantageous if, prior to step a), a product is automatically delivered and / or removed after step c), if said comparison fails positively. In this variant, goods to be manipulated are thus automatically delivered and / or transported away. For this purpose, the output of the comparison, whether a movement of the worker and the goods are in a desired range of a first movement sequence, evaluated. If the result is positive, then the current sequence of movements can be considered complete, and processed goods can be transported away and / or new goods to be processed can be delivered. Advantageously, this process does not need to be explicitly requested by the worker, for example by pressing a corresponding key, but is done automatically. Also the execution of a certain gesture is not necessary. The worker can thus work more efficiently. Of course, it can also be provided to consciously control a delivery of goods and their removal, just by pressing a button or by making a gesture.

It is particularly advantageous if the output of a correction signal or the delivery and the removal of goods are suppressed if it is determined that the worker has moved away from his workplace / work area. In this variant, therefore, the workflow is stopped when it is detected that the worker has moved away from his workplace / work area, for example, to visit the toilet. Also, the worker need not make an explicit statement to stop or restart the workflow. Stopping or starting up occurs automatically as a consequence of the worker being or not present in the workplace / workspace. Of course, can also be provided here, a delivery of goods and their removal to consciously control, as mentioned, for example, by pressing a button or by making a gesture.

It is also expedient if, prior to step a), a loading aid is transported to a provision site on the installation and the camera detects a position and / or location of the transporting loading aid relative to the provisioning location, which may include a conveyor, and this is assigned at least one segment , In this way, loading, unloading or sorting operations can be carried out very flexibly, since for this purpose the loading aids (source box) can be transported from a (large) automated warehouse. As soon as the loading aid has reached its predetermined position, a segment is assigned to it, so that steps a) to c) can be carried out. Of course, a loading aid and multiple segments can be assigned. This is advantageous, for example, when the loading aid is subdivided and has, for example, several compartments. It is generally conceivable that the actual position and / or actual position of the loading aid is detected and used for the assignment of the at least one segment, so that the at least one segment can be well aligned with the loading aid. This can be done with the camera, for example. It is also favorable in this context, however, if the presence of the loading aid on a supply space at the plant, in particular on a conveyor for source boxes and / or conveyor for destination boxes, therefore a provisioning (tendering) for a source box ( Storage container) and / or a supply space (buffer space) for a destination box (order container) is detected by means of a sensor other than the camera, for example with a light barrier or a switch. Finally, it is also conceivable that the actual position and / or actual position of the loading aid is no longer considered, and that at least one segment is assigned to a standard position when the presence of a loading aid is detected. The delivered load carrier may generally be empty or already contain goods if it is a destination box (order container), or stocked with stock items, if it is a source box (storage container). It is possible that the at least one segment is variably defined in terms of its size and / or position and approximately to a contour of a peripheral edge of the loading aid, which is located in position and / or location on the supply location, is tuned. The definition of "a contour of a peripheral edge of the loading equipment" is to be understood as meaning hen that the contour is defined by the outer walls of the loading aid (in particular loading aid without compartment division) or the contour is defined by the partitions or separation and outer walls of the loading aid (load support with compartment division or individual compartments). This is advantageous since the requirement for the positioning and / or positional accuracy of the loading aids at the provisioning location, in particular a destination box at a provisioning station for destination boxes, can be kept low and nevertheless an exact detection of a work process (eg removal of a product from a source box and storage of this product in a destination box) even with a deviation of the loading equipment from a desired position and / or desired position is possible. It has been found in practice that the actual position and / or actual position for the loading aids, which are provided at the provisioning (delivery or buffer space) succession - preferably automated - often from a desired position and / or desired Position for loading equipment is different. Thus, a loading aid may be provided sufficiently accurately in its conveying direction and / or transversely to the conveying direction on the conveyor, however, this loading aid may be rotated at the provisioning. Due to the "dynamic adaptability", the segment can be changed individually and in accordance with the current situation and the detection range of the camera can be adjusted. <br/><br/> The detection of the actual position and / or actual position preferably takes place by means of the already existing camera, but it would also be conceivable in that an additional sensor system is provided on the system which is suitable for detecting the actual position and / or actual position.

According to another embodiment of the invention, it is also possible that the at least one segment is tuned in its size dimension to a defined by the contour of a peripheral edge of the loading means surface area, wherein a surface area of the segment at least 5% larger than the surface area of the loading aid and preset becomes. The definition of "a contour of a circumferential edge of the loading aid" is understood to mean that the contour is defined by the outer walls of the loading aid (in particular loading aid without compartment division) or the contour is defined by the partitions or separation and outer walls of the loading aid (loading aids According to this variant, a "static adaptability" of the segment is possible, for example, individually to different types of loading aids, wherein the detection range of the camera is preset to different types of loading aids. A combination of "static adaptability" and "dynamic adaptability" of the segment would also be conceivable. It is also favorable if, when the comparison is positive in step c), a loading aid is transported away and an assignment of the at least one segment is resolved. In this way, system resources can be used economically and accumulation of segments can be avoided. The removed load support may in turn contain goods or be empty.

It is also advantageous if a movement sequence contains only the starting point and end point of a movement. This makes it possible to keep the calculation effort for the said method low. For the successful execution of a movement sequence, it is sufficient that the starting point and then the end point is reached by the worker (or his hand) and / or a product.

It is particularly advantageous if the detection of movements takes place in addition to a two-dimensionally imaging camera by a spatial depth sensor. This makes it possible to easily perform movements that are normal to the image plane of the camera

Way to capture. This is for example advantageous if the camera is positioned frontally to the worker and goods are rearranged between successive loading equipment. To detect the depth of space, for example, ultrasonic sensors or laser scanners can be used. A relatively new development are so-called "TOF cameras" ("time of flight"), which can measure distances with the aid of the transit time method. For this purpose, the observed area is illuminated by means of a light pulse, and the camera measures for each pixel, the time that the light needs to the object and back again. The time required is directly proportional to the distance. In this case, the function of the (conventional 2D) camera and the space depth sensor in a device, namely the TOF camera, united. In addition, it is noted that a spatial image of a workplace / workspace can of course also be obtained with the help of several 2D cameras. In general, this requires at least three 2D cameras.

It should also be noted that a "segment" in the context of the invention is not necessarily two-dimensional, but may also be three-dimensional.This three-dimensional segment can be understood as any shaped part of the three-dimensional "detection space" observed by the assistance system. While a two-dimensional segment is bounded by an envelope, a three-dimensional segment is separated from an envelope. surrounded area. It is noted that these terms are used synonymously in the context of the application. This means that a doctrine, which refers to an envelope, can be applied mutatis mutandis to enveloping surfaces and vice versa. Likewise, a teaching disclosed for 2D segments can be applied analogously to 3D segments and vice versa. For example, a segment may be classified as "occupied" if the envelope or envelope of a segment is breached by a worker, particularly his or her hand or a tool held by him / her For the sake of completeness, it will be understood that a three-dimensional segment would also be over Finally, it is also mentioned that the occupancy of a three-dimensional segment can be determined both by calculations in three-dimensional space and by calculations in several two-dimensional views. occupied "when breaking the envelopes of several projections in different planes. In particular, the projection planes or the axes of the observing cameras are orthogonal to one another. With lower requirements for detection reliability, observation from two different directions may be sufficient. It is also particularly advantageous if only a part of the detection range of the at least one camera or of the room depth sensor is subdivided into the mentioned segments. In this way, the calculation effort for the said method can be further reduced by concentrating the segmentation on an area in which the predetermined motion sequences are executed. It is conceivable, in particular, that the said part of the detection area is adapted to the respective movement sequence to be executed, that is to say that a different part of the detection area is segmented per movement sequence.

It is also particularly advantageous if the height of the worker is detected and a movement sequence is adjusted according to the detected height. This makes it possible to design the movement with respect to ergonomic aspects and thus to increase the performance of the worker and / or to avoid long-term health damage due to poor ergonomics. For example, an arm length of the worker can be taken into account. For example, if it is the job of the worker to recycle goods between different load carriers, then a worker with a short arm length can be instructed to place the load carriers closer than a worker with a long one Arm's length. By means of this active specification movements of the worker can be avoided whose harmfulness the worker himself is not aware of, for example because the individual movement causes no pain and damage occurs only after a long time and frequent execution of the harmful movement. An adaptation of a movement sequence corresponding to the detected height can also be understood to mean a distribution of the work within a group of workers. For example, activities requiring large arm length and / or high body height, such as operating a relatively tall rack, are given to workers who meet these criteria. Accordingly, small workers are entrusted with activities that are suitable for them. It is advantageous if a group of workers has different characteristics (diversification).

It is also particularly advantageous if the body size of the worker is detected and / or an area from which goods are removed and / or an area in which goods are stored is / are adjusted in accordance with the detected body size. In this way, it is also possible to automatically design the workplace / work area for the worker with regard to ergonomic aspects and thus to increase the efficiency of the worker and / or to avoid long-term damage due to poor ergonomics. In particular, the said adjustment is an adaptation of the height of said areas to understand. For example, a work surface (a work table) or conveyor for the delivery and / or removal of goods can be adjusted in height. For example, linear motors (such as spindle drives) can be used for this purpose. Of course, other drives are fully usable. Preferably, the height of a stack of goods and / or a loading aid can also be taken into account. For example, the area from which stacked goods are removed may be successively shifted upwards, so that the goods can always be taken out at the same height. Similarly, the area in which goods are stored, successively moved down so that the goods can always be stored at the same height. If goods are to be removed from or deposited in a high loading aid, this can be moved further downwards than a low loading aid, so that the upper edge of the loading aid is always at optimum height. Of course, the optimum working height can also be related to the bottom of the loading aid or to the upper edge of a stack therein. In particular, in the latter case, the height of the stack in a load carrier may vary depending on the position of the observed the camera is not visible. In this case it can be provided that the height of the stack is calculated on the basis of the height of the manipulated objects.

The adaptation of the removal area and / or the depositing area to the height of the worker is by no means limited to its height. For example, an arm length of the worker can also be used to move the removal area and / or the deposition area laterally toward the worker or away from the same. The considerations used for height adjustment are to be applied mutatis mutandis. For example, the removal area and / or the deposit area can be moved laterally sideways when goods are lined up laterally, ie, as it were, a "horizontal stack" is formed.

It is also favorable if the at least one camera is positioned in front of and / or above a workstation / workspace intended for the worker. In this way, the movements of the worker and the manipulation of the goods can be tracked very well.

It is also advantageous if the camera is movable. As a result, a camera can be used more flexibly since, as needed, it can observe a workstation / work area from different angles or can also be used to monitor several workstations. It is also conceivable, for example, for several cameras to be arranged at a workstation / workspace or to be pulled together there when a particularly difficult process is to be observed. In general, the cameras can be moved by any means. For example, they can be mounted on articulated and motorized arms ("robot arms") or moved along a rail system.

For a better understanding of the invention, this will be explained in more detail with reference to the following figures. Show it:

Fig. 1 shows a first schematically illustrated example of a system for the manipulation of

Goods in oblique view; Figure 2 is a second schematically illustrated example of a system for manipulating goods in an oblique view.

Fig. 3 shows an example of an image captured by a camera with a drawn

Motion sequence;

Fig. 4 is an example of an image captured by a camera which is segmented;

5 shows an example of an image captured by a camera with the start and end points of a motion sequence drawn in;

6 shows an example of an image captured by a camera, which is segmented only in regions; FIG. 7, like FIG. 2, only with segments / envelopes drawn in; FIG.

FIG. 8 shows the scene from above in the visual area shown in FIG. 7; FIG.

Fig. 9 as Fig. 8, only as an image of the spatial depth;

FIG. 10, like FIG. 9, only with a hand of the worker lowered into the loading aid; FIG.

11-13 is an exemplary flowchart for assisting a worker in the

Manipulation of goods and

Fig. 14 is an exemplary flowchart for monitoring "forbidden" segments.

By way of introduction, it should be noted that in the differently described embodiments, the same parts are provided with the same reference numerals or the same component designations, wherein the disclosures contained in the entire description can be mutatis mutandis to the same parts with the same reference numerals or component names. Also, the position information selected in the description, such as top, bottom, side, etc. are related to the immediately described and illustrated figure and are in the case of a change of position, to be transferred accordingly to the new situation. Furthermore, individual features or combinations of features from the illustrated and described different embodiments may represent for themselves, inventive or inventive solutions.

Fig. 1 shows a worker 1 in a plant 2 for manipulating goods 3. The plant 2 comprises a table 4, troughs in which loading aids 5 (eg boxes or boxes) are set, display devices 6 (eg lamps or LEDs), input keys 7, a screen 8 and a camera 9. The camera 9 serves for detecting movements of the worker 1 and optionally of movements of the goods 3, which is manipulated by the worker 1. In addition, a transverse to the table 4 extending conveyor 10 is provided. Finally, the system 2 also comprises means for comparing whether the detected movements of the worker 1 and, where appropriate, the product 3 are in a desired range of a movement sequence, and for outputting a correction signal, if this is not the case, wherein a motion sequence detecting and storing at least one product 3 includes. By way of example, the said means can be formed by a computer integrated in the screen 8, in which appropriate software runs.

In this way, a method for assisting a worker 1 in a device 2 for manipulating goods 3 can be realized, comprising the steps:

a) detecting movements of the worker 1 with at least one camera mounted on the plant 9,

b) comparing whether the detected movements of the worker 1 are in a desired range of a first movement sequence, and

c) outputting a correction signal and continuing at step a) with the first movement sequence if the comparison is negative, or continuing at step a) with a second movement sequence if the comparison is positive, wherein a movement sequence comprises detecting and discarding at least one product 3 includes. For example, the task of the worker 1 may be to pick goods 3 for an outgoing order. In this case, for example, goods 3 are to be taken from the left and middle boxes according to the specification of a picking system and placed in the right box 5. For example, 8 can be output on the screen: "Please place two items from the left box in the right box, then place three items from the middle box in the right box." Of course, the output can also be acoustical.Additionally or alternatively, the display devices 6 can also be used. For example, the instruction may read "Please put two items from the box marked in green into the box marked in red". Synchronously with the output, the corresponding display devices 6 are activated. Before step a), in this variant, therefore, the starting point and the end point of a movement sequence are displayed.

It would also be conceivable, for example, for the green display device 6 to flash twice to signal that two objects are to be removed. A separate text statement can then be omitted.

In addition to color markings, the starting point and the end point of the movement sequence can also be marked with different shapes, for example with a circle and a cross, or else with the number / type of goods 3 to be picked / put. In general, laser projectors and data glasses are also available for optical marking of start and end points.

Only when this process has been carried out will the next instruction be sent to the worker 1, in our case "Put three goods from the box marked in green into the box marked in red." According to the prior art, the worker 1 must have the first Confirm the process to get the next instruction, which requires an additional work step and is prone to error, as the worker 1 can in principle confirm the operation even if he did not execute it or failed.

On the one hand, the mentioned method can be omitted by the presented method, on the other hand errors of the worker 1 are recognized. In particular, the system can avoid a looming error in the approach. If, for example, in the first instruction "Please insert two items from the box marked in green into the box marked in red", it is recognized that the operator 1 removes an item from the box 5 marked in red, then a warning can be issued to the worker 1 before he has actually sorted the goods 3 incorrectly, and a warning may be issued if he does remove the goods 3 from the correct one Box 5 removes, but then moves it in the wrong direction. On the other hand, if he executes the procedure correctly, the system jumps to the next instruction.

Of course, the active confirmation of a completed work by the worker 1 is not excluded in principle. In addition to pressing appropriate keys and the execution of a designated gesture may be provided. For example, placing both hands at the left and right ends of the table edge may mean "operation completed - expect next operation." The system may also serve to sort goods 3 into the boxes 5. For example, in the right container 5, there are mixed goods 3, sorted into the middle box 5 and the left box 5 (for example, because several storage containers have fallen from awkwardness to the ground and to keep the streets running in the warehouse hastily filled into a single container) render valuable services, since the system recognizes which goods 3 from which source box 5 is placed in which destination box 5 and checks whether this is also correct.

In addition to the simple instructions given to the worker 1, the system can also provide more complex instructions. For example, a worker 1 may be instructed to rotate a detected item to a particular location before it is discarded.

In an advantageous variant of the method, a product 3 is automatically delivered before step a) and / or transported away after step c), if said comparison, whether the detected movements of the worker 1 and optionally the product 3 in a desired range of a first movement sequence lie, positive turns out. For example, goods 3 to be manipulated can be delivered and / or removed via the conveyor 10. For example, goods 3 may be picked from the boxes 5 in shipping cartons (not shown) on the conveyor 10. When the job is completed, the system automatically recognizes this and the full shipping carton is sent to the shipping department and a new empty shipping carton is transported to the next job. But the reverse case is also conceivable, namely that goods 3 are transported via the conveyor 10 from a warehouse and are picked into the waiting boxes 5. Is the corresponding number of goods 3 from the storage container ter (not shown) removed, it is transported via the conveyor 10 back to the camp.

Of course, instead of automatic delivery and automatic removal, provision may also be made for intentionally controlling a delivery of goods 3 and their removal, for example by pressing a button 7 or by making a gesture. For example, pivoting the arm to the right may mean that the conveyor 10 is transporting the next container. In a preferred variant of the method explained with reference to FIG. 1, the output of a correction signal or the delivery and removal of goods 3 is suppressed if it is determined that the worker 1 has moved away from his workstation / workspace. Thus, the workflow is stopped when it is detected that the worker 1 has moved away from his workplace / work area, for example, to go to the bathroom. The worker 1 does not need to make an explicit statement in order to stop or restart the workflow. Stopping or starting up occurs automatically as a consequence of worker 1 being or not present at the workplace / work area. Of course, it can also be deviatingly provided to intentionally control a delivery of goods 3 and their removal, for example, again by pressing a button 7 or by executing a gesture.

Advantageously, the camera 9 or an image processing software for the captured by the camera 9 image capture the structure of the worker 1, that is, for example, the torso, limbs, head and in particular hands capture. It is formed as it were a skeleton of the worker 1, whereby its movements can be detected particularly accurate or differentiated. This makes it possible to detect whether a product 3 has actually been registered. If, for example, a product 3 is only covered by the head or hull, it can be assumed that it has not been recorded. If, on the other hand, it is covered by one hand or both hands of the worker 1, then the probability is very high that the goods 3 were actually recorded. In addition, the "skeleton" of the worker 1 can be extended by tools held by him (eg pliers, grippers, etc.) and a product 3 can then be classified as captured when the goods 3 are covered by the tool becomes. Even then, the likelihood is very high that the goods 3 were actually recorded.

2 now shows a system 2 for manipulating goods 3, which is very similar to the arrangement shown in FIG. In contrast, the camera 9 but not on one

Pillar but mounted on a ceiling. In particular, it can also be provided that it is movably mounted, for example, in a rail and driven by a motor. In this way, cameras 9 can be used more flexibly. Since these move along defined paths, their location is much easier than, for example, the location of helmet cameras carried by workers 1.

In addition, the system 2 shown in FIG. 2 also comprises a room depth sensor 11.

As a result, it is possible to detect movements that are normal to the image plane of the (two-dimensionally imaging) camera 9 in a simple manner. This is advantageous, for example, if the camera 9 is positioned frontally to the worker 1 and goods 3 are rearranged between successive loading aids. Infrared sensors, ultrasound sensors or laser scanners can be used, for example, to record the spatial depth. Of course, the camera 9 and the depth sensor 11 may also be arranged in a common housing. Alternatively or additionally, so-called "TOF cameras" ("time of flight") can also be used.

Fig. 3 shows an exemplary image captured by a camera 9 (Note: the image does not show the arrangements shown in Figures 1 and 2 but another exemplary arrangement.) In particular, the boxes 5 are not arranged in special troughs, but rather - placed on the table 4). In Fig. 3, an expected motion sequence 12 is shown. The worker 1 is to put the goods / object 3 from the upper left box 5 into the lower left box 5. With the help of the camera 9 and a corresponding image processing is now checked whether the hand of the worker 1 and the object 3 are moved along the path shown. If so, then this is considered successful in performing the requested action. If this is not true, then an error signal is output and attempts to bring about the requested action. It is advantageous if in addition to the movement of the worker 1, the movements of the goods 3, which is manipulated by the worker 1, are detected before the step b) with the at least one camera 9 and the space depth sensor 11 and compared in step b), Whether the detected movements of the worker 1 and the goods 3 are within a desired range of a first movement sequence 12. The fact that not only the worker 1 is detected, but also manipulated by him goods 3, the monitoring of movement is particularly safe. In this variant, it is not sufficient, for example, merely to move the hand in a predetermined manner. Instead, the goods 3 must be moved in a predetermined manner.

It is particularly advantageous if a product 3 is classified as being detected by the worker 1 if, between the detection and removal, in particular during the entire movement sequence 12, at least partially by a hand of the worker 1 or a tool held by him (eg pliers, Gripper, etc.) is covered. As a rule, a product 3-unlike in FIG. 3 -is at least partially hidden when it is held by the worker 1. If it is not completely recognized between the detection (starting point of the movement sequence 12) and the deposition (end point of the movement sequence 12) or during the entire movement sequence 12, then it can be assumed with high probability that it is moved by the worker 12 in the predetermined manner and not, for example, lies only on a work table 4 between start and end point.

To determine whether a product 3 is at least partially hidden, for example, data from a database in which 3D models of the goods 3 are stored can be used. If a view of the relevant 3D model does not completely correspond to the view of the goods 3 detected by the camera 9 and / or the depth sensor 11, and in particular corresponds to the hidden portion (at least partially) of a hand of the worker 1 or of a tool held by him the goods 3 can be classified as recorded.

Optionally, it may also be provided that the worker 1 has to turn the product 3 into a defined position relative to the camera 9 after picking up, so that it can be recognized, for example, on the basis of its contour, coloring, etc. Fig. 4 shows the same scene as Fig. 3. Now, however, the image captured by the camera 9 is subdivided with a raster 13 into a plurality of arrays 100..146 arranged like a matrix. The detection range of the at least one camera 9 or the space depth sensor 11 is thus subdivided into segments 100..146, and a motion sequence 12 contains an occupancy of the segments 100..146 by the worker 1 and / or the goods 3. According to the motion sequence 12 the worker 1 put the object 3 from the upper left box 5 into the upper right box 5. The calculation effort is kept low in this variant of the method, since it only has to be checked whether the hand of the worker 1 and the goods 3 occupy the segments 132, 133 and 134 in the stated sequence.

In particular, if the boxes 5 can be set up randomly on a worktable 4 (that is, if, for example, predetermined troughs are missing), it may be advantageous to divide the table 4 into a predetermined grid in which the boxes 5 are to be placed. For this purpose, for example, lines can be painted on the table 4 or projected onto it. In this way it is avoided that boxes 5 are placed in an unfavorable position, for example at the crossing points of several grid lines. For example, in the illustration of FIG. 4, the rear boxes 5 from the perspective of the worker 1 are relatively unfavorable because they occupy comparatively many segments.

It is also conceivable that a motion sequence 12 contains only the starting point 132 and end point 134 of a movement, as shown in FIG. 5. In this case, it is only checked whether the hand of the worker 1 and the goods 3 occupy the segments 132, and 134 in the order named. This makes it possible to keep the calculation effort for the said method even lower. For the successful execution of a movement sequence, it is sufficient that starting point and then the end point by the worker 1 (or his hand) and / or a product 3 is achieved.

FIG. 6 shows a further variant of the method, in which only part of the detection range of the at least one camera 9 or of the room depth sensor 11 is subdivided into the mentioned segments 100... By concentrating the segmentation on an area in which the predetermined motion sequences are executed, the calculation effort for the said method can be reduced even further. In the application shown in Fig. 6 are used as an example only the boxes 5 associated segments 100..103. Accordingly, the check can be limited to whether the expected movement sequence 12 is occupied by occupying the current motion sequence 12 required, ie, "active" segments 100 and 103 or optionally also by non-occupancy of not required for the current motion sequence 12, the is called "inactive" segments 101 and 102 has been executed. It is also conceivable, of course, that several segments are provided in the action area of the worker 1.

It is also conceivable, in particular, for the segmented part of the detection area to be adapted to the movement sequence 12 to be executed in each case, that is to say for each movement sequence 12 a different part of the detection area to be segmented. For the motion sequence 12 shown concretely, this would mean that the detection area is subdivided only into the segments 100 and 103 (or also into further segments required in the operating range of the worker 1).

Thus, in the embodiment presented, an "envelope curve" is placed around a respective box 5. A segment 100 and 103 can be classified as occupied if this envelope is breached by the worker 1, in particular his hand In order to avoid accidental penetration of the envelope, it can also be provided that a segment 100 and 103 must be occupied for a certain time (eg one second) so that it is also classified as occupied.

It is also conceivable that, prior to step a), a loading aid 5 (eg box, container) is preferably transported in an automated manner and this is assigned at least one segment 100 and 103. In this way, loading, unloading or sorting operations can be carried out very flexibly, since for this purpose the loading aids 5 can be transported from a (large) warehouse. As soon as the loading aid 5 has reached its predetermined position on a provisioning station at the installation 2, in particular a provisioning station (delivery point) for a source box (storage container) and / or a provisioning station (buffer space) for a destination box (order container) , this is assigned a segment 100 and 103. Of course, a loading aid 5 can also be assigned a plurality of segments 100 and 103, in particular if the loading aid 5 is subdivided. It is generally conceivable that the actual position and / or actual position of the loading aid 5 is detected and used for the assignment of the at least one segment 100 and 103, so that the at least one segment 100 and 103 and the loading aid 5 are coordinated with each other can be done for example with the help of the camera 9. It is also favorable in this context, however, if the presence of the loading aid 5 is detected at the place of supply to the system 2 by means of a sensor other than the camera 9, for example with a light barrier or a switch. Finally, it is also conceivable that the actual position and / or actual position of the loading aid 5 is disregarded, and that at least one segment 100 and 103 is assigned to a standard position when the presence of a loading aid 5 is detected. The delivered loading aid 5 may generally be empty or already contain goods 3.

It is also favorable if, with a positive outcome of the comparison in step c), a loading aid 5 is transported away and an assignment of the at least one segment 100 and 103 is resolved. In this way, system resources can be used economically and accumulation of segments 100 and 103 can be avoided. The transported offloading means 5 may in turn contain goods 3 or be empty.

FIG. 7 now shows the workplace already shown in FIG. 2, but now with segments or envelopes / enveloping surfaces 100... 102 above the boxes 5. This is intended to visualize that the segments / envelopes 100... 102 are not necessarily planar or two-dimensional but can also be spatial or three-dimensional. This is particularly advantageous when the workstation / workspace can be detected three-dimensionally with a room depth sensor 11 or an arrangement of several 2D cameras. Of course, the teaching on two-dimensional segments / envelopes 100..146 disclosed above is also applicable without restriction to three-dimensional segments / envelopes 100... 102. In addition, it is mentioned that the segments / envelopes 100... 102 in FIG. 7 can of course also have a different shape, for example the shape of a hemisphere or even irregular shapes.

Fig. 8 shows the scenery shown in Fig. 7, now in fragmentary form from above. Specifically, the table 4, the loading equipment 5, the goods 3, the display devices 6, the input keys 7 and a hand of the worker 1 are shown, which differ from the Fig. 7 all- Now it is located above the right loading aid 5. The loading aids 5 are in turn assigned to the segments 100..102. FIG. 8 shows the scene in the visual range, that is to say as it is perceived by the human eye. FIG. 9 now shows the same scene as an image of the spatial depth, that is to say as perceived by a room-depth camera or a room depth sensor 11. Different areas of the room depth camera or the depth sensor 11 different areas are hatched differently in Fig. 9. In a real version, different distances can of course also be assigned to different gray values or color values. For example, more distant objects are light gray, closer objects are shown in dark gray.

In FIG. 9, it can be seen from the hatching that the hand of the worker 1 is positioned above the loading aid 5, since the hand is hatched differently than the bottom of the loading means 5 and also unlike the table 4 9 can be seen that the display devices 6 and the input keys 7 "disappear" because they are just installed in the concrete example shown in the table 4.

In FIG. 10, the hand of the worker 1 is lowered relative to the illustration shown in FIG. The hand now has the same hatching as the goods 3. Fig. 10 shows the

Arrangement thereby simplifies, since the entire hand is on a different level as the arm of the worker 1. In reality, the arm and the hand of course, are inclined and not stepped. For example, a segment 100..102 is considered occupied when the worker's hand 1

(and / or for example a tool held by the latter) in a specific spatial area is detected. In the specific case, it is assumed that this spatial area corresponds to the area covered by a loading aid 5. In Fig. 9, the hand of the worker 1 is positioned above this area. The segment 100 is therefore considered not occupied. In FIG. 10, by contrast, the hand of the worker 1 is lowered into the loading aid 5. The

Segment 100 is therefore considered occupied. In this way it is avoided that a hand detected only above a segment 100... 102 detects that the system is detecting or storing it a product 3 is misunderstood. Only the occupation of a (spatial) segment 100..102 is regarded as detecting or storing a product 3.

In the above example, it has been assumed that the space area in which the detection or deposition of a commodity 3 is detected corresponds to the area encompassed by a loading aid 5. This is not a necessary condition. The said spatial area can also comprise only a partial area, in particular from the floor, and thus correspond to the area which would occupy a liquid filled in the loading aid 5. However, the space area can also be arranged at a distance from the floor, similar to the spatial area that an oil layer floating on water would occupy in the loading aid 5. However, the spatial area can also protrude beyond the loading aid 5 (cf. FIG. 7) or virtually hover above the loading aid 5.

From FIGS. 8 to 9 it is clear that the detection of the spatial depth is sufficient for the method of assisting the worker 1 per se. The camera 9 can therefore be designed as a room-depth camera and be provided in particular exclusively for detecting the spatial depth. A separate room depth sensor 11 is then no longer required. In general, the boundaries between the room-depth camera and the room depth sensor are fluid and, in principle, the variants explained in connection with FIGS. 1 to 7 and the resulting advantages are analogously applicable to an embodiment in which the camera 9 is provided exclusively for detecting the room depth.

Preferably, a room depth camera 9 is positioned directly above the worker 1 so that it directly provides the image shown in FIG. 9 and FIG. 10, respectively. It is also conceivable, however, that this is positioned slightly above the worker 1, taking into account that e.g. the loading aids 5 and also the segments 100..102 are not equidistant from the room-depth camera 9 due to their distance in the plane. The segments 100..102 are then considered to be occupied at different distances, which are also determined by the position of the respective segments 100... 102 in the plane.

It would also be conceivable to correct the image captured by the spatial depth camera 9 insofar as the relative position of a segment 100... 102 to the spatial depth camera 9 in the plane is taken into account. The "image plane" of the spatial depth camera 9 can therefore be inclined mathematically. Despite the position of the room-depth camera 9 obliquely above the worker 1, the segments 100..102 then appear equidistant.

Quite generally and therefore not only in connection with the embodiment illustrated in FIGS. 8 to 10, further parameters can be used to decide whether a segment 100... 102 is now occupied or not. For example, the residence time of the worker's hand 1 in a segment 100..102 can be used for the decision. In particular, if the residence time for the detection of a product 3 is shorter than would generally be required for it, then the segment 100..102 can be considered as not occupied. Another parameter for said decision may be the speed of the worker's hand 1. If a segment 100..102 passes through too quickly, this can also be an indication that a product 3 was not properly recorded or filed. Furthermore, the course of the speed and / or the acceleration of the hand of the worker 1 for the decision whether a segment 100..102 is now occupied or not, are used. If a segment 100..102 is crossed at a constant speed or no or only slight accelerations are measured, this can also be an indication that a product 3 has not been recorded or stored properly. Usually, a hand 3 is slowed down when detecting or storing a product. Conversely, a detection of such a delay can again be a positive indication for the detection / removal of a product 3. In particular, when reaching into a loading aid 5, a reversal of the movement in the vertical direction should also be detectable. If this is not the case, there may be an error in the evaluation. For detection of a hand or a tool in or over a segment 100... 102 it may also be provided that simply a region of a certain size (for example a certain number of pixels) in the relevant one

Image area is detected. That is, in a simple embodiment, it is only checked if an area of a certain (minimum) size is detected. Additionally or alternatively, the shape of the recognized area can also be evaluated. For the correct execution of a work step, it may also be provided that the starting and ending segments in which goods 3 are picked up or stored, but certain intervening segments are not occupied, but only swept over. If, for example, a product 3 from the right load support 5 in the relocate the left loading equipment 5, it can be provided that the first segment 100 must be occupied (picking up the goods 3), then the segment 101 must be swept over, ie the hand must be passed over it, and finally the segment 102 must be occupied (Store the goods 3).

Conversely, it is conceivable that "forbidden" segments 100..102 are defined. If it is required, for example, to relocate a product 3 from the middle loading aid 5 into the left loading aid 5, then it can be provided that the segment 100 may not be swept over, but exclusively the segments 101 (picking up the goods 3) and the segment 102. must be placed (storage of goods 3).

In a further preferred and generally applicable embodiment, the body size of the worker 1 is detected and a movement sequence 12 adapted according to the detected body size.

This makes it possible to design the movement sequence in terms of ergonomic aspects and thus to increase the performance of the worker 1 and / or to avoid long-term health damage due to poor ergonomics. For example, an arm length of the worker 1 can be taken into account. If, as shown in FIG. 6, the task of the worker 1 is to reorder goods 3 between different loading aids 5, then a worker 1 with a short arm length can be instructed to place the loading aids 5 closer than a worker 1 with a long arm length. This specification can be done, for example, via the screen 8 or implicitly also in that a grid projected onto the table 4 is narrower or wider. In this way, an area from which goods 3 are removed and / or an area where goods 3 are stored are adjusted according to the detected body size of the worker 1.

In particular, the said adaptation also means an adjustment of the height of said areas. For example, a worktable 4 or the conveying means 10 can be designed to be height-adjustable. For example, for adjusting the height linear motors (such as spindle drives) or other drives can be used. If the presence of a large worker 1 is detected by the camera 9, then the working table 4 and / or the conveyor 10 moved up. If a small worker 1 detected, then down accordingly.

Preferably, the height of a stack of goods and / or a loading aid 5 can also be taken into account. For example, the bottom surfaces of the wells shown in Figs. 1 and 2 may be height adjustable. If goods 3 can be removed from or deposited in a high loading aid 5, this can be moved further downwards than a low loading aid 5, so that the upper edge or the bottom of the loading aid 5 is always at optimum height.

In particular, when goods 3 are to be stacked, for example, garments, the area from which stacked goods 3 are removed may be successively shifted upward, so that the goods 3 are always taken out at the same height. Likewise, the area in which goods 3 are stored, successively moved down, so that the goods 3 are always stored at the same height.

This method is also applicable when goods 3 are to be stacked in a load carrier 5. If the height of the stack located in a loading aid 5 can not be seen by the camera 9, the height of the stack can also be calculated on the basis of the height of the manipulated objects.

The adaptation of the removal area and / or the deposit area to the height of the worker 1 is by no means limited to the height. For example, an arm length of the worker 1 can be used to move the removal area and / or the deposit area accordingly laterally to the worker 1 or away from it. The considerations used for height adjustment are to be applied mutatis mutandis. For example, the removal area and / or the deposit area can be successively moved laterally when goods 3 are lined up laterally, so as it were a "horizontal stack" is formed.

Through these active defaults movements of the worker 1 can be avoided, the harmfulness of which the worker 1 is not aware of, for example because the individual movement is not Pain causes and damage only after a long time and frequent execution of the harmful movement occur.

An adaptation of a movement sequence 12 corresponding to the detected body size can also be understood as a distribution of the work within a group of workers 1. For example, activities requiring large arm length and / or high body height, such as operating a relatively tall rack, are awarded to workers 1 who meet these criteria. Accordingly, small workers 1 are entrusted with activities that are suitable for them. It is advantageous if a group of workers 1 has different characteristics (diversification).

In this context, it is generally pointed out that the method presented is, of course, not limited to use on a worktable 4 but can also be applied to other devices in a warehouse, such as working on a shelf. For example, such shelves can serve not only for storing goods 3, but also for picking. Notwithstanding the arrangements in FIGS. 1 and 2, it is of course also possible to use a plurality of cameras 9 and / or room depth sensors 11 whose detection ranges can also overlap. In particular, a camera 9 and / or a room depth sensor 11 can also be designed to be movable.

By detecting both the movement of a worker 1 and the goods 3 moved by him from a remote observer site, the system can be used very flexibly under changing conditions and recognizes the monitored workflow with high reliability. This significantly improves the monitoring of a workflow. In particular, the following results in particular the advantages explained in detail above:

The movements of a worker 1 can be monitored independently of the viewing direction.

The worker 1 is more efficient because he is not affected by a helmet camera.

The system can not fall "out of step" since the position of the observing camera 9 is always known.

Monitoring the manipulation of hard-to-distinguish items becomes facilitated.

The supervision of several workers 1 who work together and, for example, hand over objects or carry an object together is facilitated.

Restrictions or preferences of a worker 1 given by his body structure may be taken into account.

FIGS. 11 to 13 now show an exemplary procedure for assisting a worker 1, in which, after placing an order on it, the presence of a loading aid 5 (LHM), in particular a source box and / or a destination box, is checked. If this is present, then its position and / or position is detected and the loading aid 5 accordingly a segment 100..146 assigned. In the present example it is assumed that this is an envelope surface. Of course, it is also possible that this is an envelope volume. As mentioned above, a "static adaptability" or "dynamic adaptability" is possible for the segment 100..146.

Thereafter, an object search is started in which it is examined whether a worker's hand 1, a tool held by him, or the like is within a predetermined range. First, it is checked whether a detected object has or exceeds a certain size. If this is not the case, the object search is continued. If this is true, however, it is checked whether the object is located above a segment 100..146 which is assigned to a loading aid 5. If this is not the case, the object search is continued. However, if this is true, then it is checked whether the relevant segment is 100..146 part of the given order, that is, whether it is a expected in the given sequence of motion segment 100..146. If this is not the case, a warning is issued and the object search continues. If this is true, then it is checked whether the object penetrates the envelope of the segment 100..146 or the object is in the envelope volume. If this is not the case, the object search continues. If this is the case, however, it is checked whether the object penetrates said envelope for a predetermined time or is located in said envelope volume. If this is the case, a good 3 is classified as captured or stored (ie a "pick" or a "put" detected). If this is not the case, then the object search continues. The sequence ends when the order ends with the pick or the put. Otherwise, the object search continues.

At the same time as the sequence shown in FIGS. 11 to 13, it is also possible to check whether an object, that is to say a hand of a worker 1, a tool held by him or the like, does not lie in a predetermined area ("forbidden segment"). When placing an order, forbidden segments 100..146 are defined for this purpose, in the present example it is assumed that this is an envelope surface or an envelope volume.

The subsequent object search again checks whether a detected object has a certain minimum size. If this is not the case, the object search continues. If this is true, then it is checked whether the object is in a forbidden segment 100..146. If this is the case, a warning is issued and continued with the object search. If this is not true, then it is checked whether the job is completed. If this is not the case, the object search continues. Otherwise, the process ends.

The embodiments show possible embodiments of an assistance system according to the invention, it being noted at this point that the invention is not limited to the specifically illustrated embodiments of the same / same, but also various combinations of the individual embodiments are possible with each other and this possibility of variation due to the teaching of technical Acting by objective invention in the skill of those working in this technical field is the expert. Thus, all conceivable embodiments that are possible by combinations of individual details of the embodiment variant shown and described are also included in the scope of protection.

In particular, it should be noted that in reality the illustrated devices may also comprise more or fewer components than shown and are sometimes shown in greatly simplified form in the figures. For the sake of the order, it should finally be pointed out that the illustrated devices as well as their components have also been shown partly out of scale and / or enlarged and / or reduced in size for a better understanding of their design.

The task underlying the independent inventive solutions can be taken from the description.

REFERENCE NUMBERS

worker

Plant for the manipulation of goods

Object (goods)

table

Loading equipment (box, box)

indicator light

button

screen

camera

Conveyor

Space depth sensor

motion sequence

grid

field

Claims

Patent claims

1. A method for assisting a worker (1) in a facility (2) for manipulating goods (3), comprising the steps of:

a) detecting movements of the worker (1) with at least one camera mounted on the installation side (9),

b) comparing whether the detected movements of the worker (1) lie in a desired range of a first movement sequence (12),

c) outputting a correction signal and continuing at step a) with the first motion sequence (12) if the comparison is negative or continuing at step a) with a second motion sequence (12) if the comparison is positive, wherein a motion sequence (12) comprises detecting and depositing at least one product (3), wherein the detection range of the at least one camera (9) is divided into segments (100..146) and wherein a motion sequence (12) occupies an occupancy of the segments (100 .. 146) by the worker (1).

2. The method according to claim 1, characterized in that the movements of the product (3), which is manipulated by the worker (1), are detected before the step b) with the at least one camera (9) and compared in step b) is whether the detected movements of the worker (1) and the goods (3) are in a desired range of a first movement sequence (12).

3. The method according to claim 2, characterized in that a product (3) is classified as detected by the worker (1) if it covers at least partially between the detection and depositing by a hand of the worker (1) or a tool held by him is.

4. The method according to any one of claims 1 to 3, characterized in that before step a) at least the starting point (132) and / or the end point (134) of a movement sequence (12) are displayed.

5. The method according to any one of claims 1 to 4, characterized in that before

Step a) a product (3) automatically delivered and / or removed after step c), if said comparison is positive.

6. The method according to any one of claims 1 to 5, characterized in that the

Output of a correction signal or the delivery and removal of goods (3) is suppressed, if it is determined that the worker (1) has moved away from his workplace.

7. The method according to any one of claims 1 to 6, characterized in that prior to step a) a loading aid (5) transported to a supply point at the plant (2) and preferably the camera (9) a position and / or location of the transported Loading aid (5) detected relative to the provisioning and assigned to this at least one segment (100..146).

8. The method according to any one of claims 1 to 7, characterized in that the at least one segment (100..146) defined variable in its size and / or position and approximately to a contour of at least one peripheral edge of the loading aid (5), which is in position and / or position at the deployment location, is tuned.

9. The method according to any one of claims 1 to 8, characterized in that the at least one segment (100..146) in its size dimension on a by the contour of at least one circumferential edge of the loading aid (5) defined surface extent is tuned, wherein a surface area of the segment (100..146) is at least 5% larger than the surface area of the loading aid (5) and preset.

10. The method according to any one of claims 7 to 9, characterized in that a presence of the loading aid (5) is detected by means of a sensor different from the camera (9).

11. The method according to any one of claims 1 to 10, characterized in that at a positive outcome of the comparison in step c) a loading aid (5) transported away and an assignment of the at least one segment (100..146) is resolved.

12. The method according to any one of claims 1 to 11, characterized in that a movement sequence (12) contains only the starting point (132) and end point (134) of a movement.

13. The method according to any one of claims 1 to 12, characterized in that the detection of movements in addition to a two-dimensional imaging camera (9) by a space depth sensor (11).

14. The method according to any one of claims 1 to 13, characterized in that only a part of the detection range of the at least one camera (9) or the spatial depth sensor (11) is divided into said segments (100..146).

15. The method according to any one of claims 1 to 14, characterized in that the body size of the worker (1) detected and a movement sequence (12) is adjusted according to the detected body size.

16. The method according to any one of claims 1 to 15, characterized in that the body size of the worker (1) detected and an area from which goods (3) are removed, and / or an area in which goods (3) are stored , is adjusted according to the recorded height.

17. The method according to any one of claims 1 to 16, characterized in that the at least one camera (9) is positioned in front of and / or above a for the worker (1) provided workplace.

Assistance system for assisting a worker (1) in a facility (2)

Manipulation of goods (3), comprising:

at least one camera mounted on the installation side (9) for detecting movements of the worker (1), Means for comparing whether said movements of the worker (1) lie within a desired range of a movement sequence (12), and for outputting a correction signal if this is not the case, a movement sequence (12) detecting and discarding at least one commodity (3), wherein the detection range of the at least one camera (9) is subdivided into segments (100..146) and wherein a motion sequence (12) occupies an occupancy of the segments (100..146) by the worker (1). contains.

19. Assistance system according to claim 18, characterized in that the camera (9) for detecting movements of the goods (3), which is manipulated by the worker (1) is formed, and the comparison means for comparing whether the movements of Worker (1) and the goods (3) in a desired range of a movement sequence (12) are formed.

20. Assistance system according to claim 18 or 19, characterized in that the camera (9) is positioned in front of and / or above a for the worker (1) provided workplace.

21. Assistance system according to one of claims 18 to 19, characterized in that the camera (9) is movable.

22. Use of a method according to one of claims 1 to 17 and / or an assistance system according to one of claims 16 to 21 in a warehouse.

23. Use according to claim 22, characterized in that a movement sequence (12) relates to the picking and / or resorting of goods (3).