Ejection device and method for operating an ejection device

Abstract

The invention relates to an ejection device. This device comprises at least one actuator with a working range within which the actuator can exert a force impulse on an object. The ejection device also comprises at least one detector with a detection range. The detection range is intended to overlap with the working range of the actuator or to be located at a sufficiently small distance from it. The ejection device is configured such that a force impulse is triggered when the detector detects the presence of an object within the working range of the actuator.

Description

[0001] The invention relates to an ejection device. This device comprises at least one actuator with an effective range within which a force impulse can be exerted on an object by the actuator. The ejection device also comprises at least one detector with a detection range. The detection range should overlap with the effective range of the actuator or be at a sufficiently small distance from it. The ejection device is configured such that a force impulse is triggered when the detector detects the presence of an object within the effective range of the actuator.

[0002] In sensor-assisted bulk material sorting, the localization and characterization of individual particles in the material stream is carried out using a sensor (inspection sensor technology). According to the state of the art, line-scanning sensors, such as color line-scan cameras, are predominantly used for this purpose. Based on the characterization of a particle, a sorting decision is calculated, i.e., a decision is made as to whether or not a particle should be removed from the material stream. Pneumatic quick-acting valves are predominantly used for physical removal.

[0003] There is a time lag between sensor detection and physical separation. This is primarily due to the time required for sensor data processing, i.e., for performing localization and characterization. Therefore, based on the localization at the time of detection, a prediction must be made as to when one or more of the quick-acting valves must be activated to remove a particle from the material stream. This prediction results in the so-called blow-out window, which describes the duration and number of nozzles for activation.

[0004] Since no movement information about individual particles is available (the use of line-scanning sensors only produces a single image of the particle), global assumptions must be made. In state-of-the-art sorting systems, the entire system is set to the average transport speed, and it is assumed that there is no movement perpendicular to the conveying direction. Thus, a fixed time offset between detection and separation is assumed. The state in which this assumption applies is referred to below as optimal material settling.

[0005] For many products (especially spherical ones, e.g., peppercorns), it is very difficult, if not impossible, to achieve such optimal material stabilization. This leads to incorrect nozzle activation, with such an error being reflected in both the temporal component (too early / late activation) and the local component (incorrect nozzle(s) activated within the nozzle bar). Deviations from the average transport speed and / or the presence of transverse movements can lead to sorting errors because particles are not removed.

[0006] Ultimately, the degree of stabilization depends not only on the product itself but also on the transport mechanism used. For example, very long conveyor belts generally provide greater material stabilization than short conveyor belts or chutes. However, the use of long conveyor belts comes with disadvantages, such as high costs (purchase, space, maintenance, etc.).

[0007] As explained in Section 1, state-of-the-art sorting systems use the average transport speed and the assumption that no transverse movements exist as global parameters to determine the discharge time and location. To compensate for possible deviations from this, a large discharge window is often selected. This means that the dimensioning of the discharge window (temporal and spatial components) is not only based on the location (and possibly size) of the object determined from the sensor data, but is deliberately enlarged. Pragmatically speaking, the nozzles are activated for a longer period, e.g., shortly before the expected arrival of the object at the nozzle bar and even beyond, and neighboring nozzles are also activated if necessary. This approach has at least two disadvantages. Firstly, it increases the risk of mistakenly ejecting particles that should not be discharged.This occurs whenever a particle in the "non-reject" category is in close proximity to a particle in the "reject" category, with the tolerance regarding the proximity of the particles decreasing as the discharge window size increases. Furthermore, larger discharge windows require more compressed air. Since the generation of compressed air represents a significant portion of the operating costs of bulk material sorting, this is undesirable.

[0008] An innovative method that has been researched in recent years is the use of area scan cameras combined with object tracking in optical bulk material sorting. By using a high-speed camera, particles contained in the material stream are observed at multiple points in time and tracked via a multi-object tracking system. While tracking individual or a few objects is a comparatively simple problem, tracking numerous, closely spaced particles is an algorithmically difficult problem. For the application field of optical bulk material sorting, thousands of particles must be tracked simultaneously – and in real time. By using area scan cameras and tracking individual particles over time, motion parameters, such as the velocity components in and across the transport direction, can be determined individually for each particle.These parameters can be used with the help of a model to precisely estimate future positions. For optical bulk material sorting, this means that the actuators can be controlled much more accurately. While this has already achieved a significant improvement, as any transverse movements of the bulk material are recorded, it is still only a blow-out prediction. Depending on the observation area, effects at the discharge edge, i.e., immediately before separation, may also not be taken into account. Further complicating the situation is the fact that the use of area-scanning sensors in optical bulk material sorting is not widespread and would require a complex redesign of existing systems.

[0009] The object of the present invention is therefore to reduce the proportion of incorrectly ejected objects while at the same time reducing the technical effort and costs.

[0010] The object is achieved by the ejection device according to claim 1 and the method for operating an ejection device according to claim 14. The respective dependent claims specify advantageous developments of the ejection device according to the invention and of the method according to the invention.

[0011] According to the invention, an ejection device is provided that has at least one actuator. The actuator has an effective range within which a force impulse can be exerted on an object located within the effective range. The effective range can also be defined by its intended function. The effective range can be considered the range within which the actuator is capable of exerting a sufficiently strong force impulse on an object so that it is ejected.

[0012] The ejection device according to the invention also has at least one detector with a detection zone. The detection zone is the zone in which objects can be detected by the detector. In the detection zone, the detector can preferably interact with the objects to detect their presence.

[0013] According to the invention, the detection range overlaps partially or completely with the effective range of the actuator. The fact that the detection range overlaps with the effective range means that there is at least one spatial region that is part of both the detection range and the effective range. The detection range and the effective range can therefore be identical or have an intersection or intersection area, or the detection range can be a subset or a subset of the effective range, or the effective range can be a subset or a subset of the detection range. Alternatively, the effective range can have a distance of less than or equal to 10 mm from the detection range. If the detection range and the effective range are separated by less than or equal to 10 mm, this distance should preferably be the smallest distance between the detection range and the effective range.

[0014] According to the invention, the detection zone should be designed and arranged such that the presence of an object in the effective range of the corresponding actuator can be detected by the detector. The presence of the object in the effective range should preferably be detected directly, i.e., the detector should respond to the presence of the object as soon as the object enters the effective range of the actuator. For example, the actuator can interact with the object to detect its presence in the effective range while the object is located in the effective range. In contrast, indirect detection should be understood as a situation in which the object is detected by the detector at a specific time, and its presence in the effective range at another time is inferred based on the object's trajectory.

[0015] According to the invention, the ejection device is also configured to cause the actuator to generate a force pulse only if the detector detects the presence of an object within the effective range of this actuator. The actuator should therefore be controlled by the detector in such a way that the detection of an object by the detector within the effective range of the actuator can, but optionally does not necessarily, lead to the generation of a force pulse by the actuator. However, the actuator should preferably not generate force pulses without being triggered by the detector due to the detection of an object.

[0016] Ultimately, it is important that the detection of an object within the effective range can trigger the force impulse. In some embodiments of the invention, however, the generation of the force impulse upon detection of the object is not mandatory, for example, if the device is only intended to eject objects with certain properties that can be additionally determined. The fact that the detector detects the presence of an object within the effective range of the corresponding actuator is therefore a necessary, but not necessarily sufficient, condition for the actuator to generate a force impulse.

[0017] In an advantageous embodiment of the invention, the ejection device can be configured to receive or generate a control signal by means of which the force pulse can be enabled or disabled. The control signal can thus be, for example, an enable signal or a disable signal, as required. In this optional embodiment, this control signal can be added to the detection of the object by the detector as a further necessary condition for generating the force pulse. The control signal preferably represents a further necessary condition, but not necessarily a sufficient one together with the detection of the object in the effective range.In this embodiment, the ejection device should therefore be configured such that the detector is only triggered to generate the force impulse if the detector detects the presence of an object in the effective range of this actuator and, in addition, the control signal enables or does not block the generation of the force impulse.

[0018] Such a control signal can be used, for example, to specify a release period. For example, existing inspection sensors can be used to determine an approximate time frame in which a particle to be ejected passes through the effective range of one of the actuators. This rough and possibly generously selected time interval can then be transmitted to the actuator or an actuator controller. Within this release window, the actuator can then optionally always generate a force pulse whenever a suitable object is within the effective range of the actuator.

[0019] In an advantageous embodiment, the ejection device can have an upstream sensor that is configured to observe objects in an object stream. The upstream sensor can advantageously be configured to determine at least one property of the objects. The ejection device can then be configured to generate said control signal such that the force impulse of at least one of the actuators is released when the property of the object determined by the upstream sensor matches a predetermined property. A time offset can be provided between the determination of the at least one property by the upstream sensor and the generation of the control signal, which is equal to the movement time of the object from the time of observation by the upstream sensor to the arrival in the effective range of the corresponding actuator.For this purpose, the trajectory of the object can also be advantageously determined and / or taken into account.

[0020] In an advantageous embodiment of the invention, at least one of the detectors can be configured to detect, in addition to the presence of the object in the effective range of the corresponding actuator, at least one property of this object. The presence and property can preferably be detected simultaneously. The ejection device can then be configured to cause the actuator to generate the force impulse only if the detected property of the object also matches a predetermined property. In this embodiment, therefore, not every presence of an object in the effective range leads to the generation of a force impulse, but only the presence of objects that exhibit a predetermined property.

[0021] In an advantageous embodiment, the detector can be designed such that it detects the presence of an object depending on a property of the object, i.e. that it only detects the presence of objects that have the specified property within the scope of the detector's perception accuracy. An example of such a design could be not sorting out objects of a certain color. For this purpose, the detector could be a camera or a color sensor that is aimed at a background of the corresponding color. If an object with the color of the background (within the scope of the detector's perception accuracy) moves past between the camera and the background, it will not be detected by the camera due to its color, and no force impulse will be triggered.Objects with other colors, however (i.e. those whose color differs from the background by more than the perceptual accuracy) can be detected by the camera, so that a force impulse is generated for them.

[0022] In another example, at least one of the detectors can be configured as a camera or color sensor directed at a background of a specific brightness. The brightness of the background can be selected such that, within the perceptual accuracy of the sensor or camera, it matches the brightness of those objects that should not receive a force impulse. Here, too, the camera will not detect objects whose brightness matches the brightness of the background, within the perceptual accuracy of the camera.

[0023] Advantageously, the at least one actuator can be a nozzle, particularly preferably a compressed air nozzle and / or a water jet nozzle.

[0024] The at least one detector can advantageously comprise or be an optical sensor, a light barrier, a retro-reflective light barrier, a color detector, a photosensor, a proximity switch, a capacitive sensor, an eddy current sensor, and / or a magnetic sensor. If the detector is a light barrier, it is advantageous if it measures across the effective range of at least one of the actuators, i.e., if the light source and the light sensor of the light barrier are located on opposite sides of the effective range of the corresponding actuator. Advantageously, the effective ranges of the actuators can each be observed by a separate detector. If there are multiple actuators, a detector can be provided for each of these actuators, which monitors the effective range of the corresponding actuator.

[0025] In an advantageous embodiment, the ejection device can comprise a plurality of actuator-detector combinations. Each of the actuator-detector combinations can contain one of the actuators and one of the detectors, with the detection range of the detector overlapping with the effective range of the corresponding actuator in the manner described.

[0026] In an advantageous embodiment, a plurality of actuator-detector combinations can be arranged next to one another along a straight line. In this way, they can form, for example, an actuator bar. It is also possible for a plurality of actuator-detector combinations to be arranged next to one another along several parallel straight lines. In this way, they can form, for example, several parallel actuator bars.

[0027] The ejection device can therefore use actuator-detector combinations individually or in combination. In combination, the actuator-detector combinations can be arranged in any configuration, for example, in the form of a conventional linear nozzle bar or a non-linear nozzle bar. In principle, any number of rows of such actuator bars can be connected in series perpendicular to the transport direction of the material flow, i.e., the flow of objects.

[0028] In an advantageous embodiment, some or all of the actuator-detector combinations can have a unidirectional or bidirectional communication channel with at least one, preferably adjacent, actuator-detector combination, preferably with all immediately adjacent actuator-detector combinations. Control signals can be transmitted between the actuator-detector combinations via these communication channels. A unidirectional communication channel may be sufficient if, for example, the control signal is to be transmitted in only one direction, for example, along the transport direction of the objects.

[0029] At least some or all of the actuator-detector combinations can then be configured to cause their corresponding actuator to generate a force pulse only if they have received a control signal permitting the generation of the force pulse. Depending on the requirements, the control signal can be an enable signal that enables the generation of the force pulse or a blocking signal that blocks the generation of the force pulse.

[0030] In a particularly advantageous embodiment, at least two parallel actuator beams can be provided, each having a plurality of actuator-detector combinations. An actuator-detector combination of a first of the actuator beams and an actuator-detector combination of a second of the actuator beams can then be arranged one behind the other in a flow direction in which the objects move. The actuator beams can advantageously be arranged parallel to one another. The actuator-detector combinations that are at the rear in the flow direction can then be configured to receive a control signal from the corresponding actuator-detector combination that is at the front in the flow direction.The actuator-detector combination located at the rear in the direction of flow can then be configured to cause the actuator of this actuator-detector combination, i.e. its actuator, to generate the force impulse when its detector detects the presence of the object in the effective range of its actuator and, at the same time, the control signal enables the generation of the force impulse.

[0031] The fact that the actuator-detector combinations are arranged one behind the other in the flow direction means that an object in the object stream that passes the front actuator-detector combination then passes the rear actuator-detector combination.

[0032] Advantageously, the ejection device can have at least one control unit configured to control the at least one actuator depending on at least one signal from at least one of the detectors and / or optionally depending on the control signal. Such a control unit can be, for example, a microcontroller or a PC. Alternatively, such a control unit can also be configured in analog form, which can enable rapid processing of the detector signals. The control unit can be connected to the sensor as well as to the actuator.

[0033] A program, for example, on a microcontroller, can determine at any time whether an object is within the detection zone (and thus also within the effective range) or not. However, its functionality can also be implemented without a dedicated controller using a PC, for example, the one on which the data from an inspection sensor is evaluated. Such a PC can also control the corresponding actuator. Due to the binary decision criteria and the fact that no or very rapid additional characterization of the object within the effective range takes place, the data can be evaluated virtually in real time.

[0034] In a very simple design, the reject device can be used independently. For this purpose, the controller (or PC) can be programmed so that the actuator is always activated when an object is within the effective range. This corresponds to rejecting an object during a sorting process (without characterization or with characterization enabled by the detector).

[0035] The invention also provides a method for operating an ejection device as described above. At least one object moves through the effective range of at least one of the actuators. The presence of the object in the effective range of the actuator is detected by at least one of the detectors. The actuator generates a force impulse only if the detector detects the presence of the object in the effective range of this actuator.

[0036] The method can be particularly advantageously implemented in an ejection device comprising multiple actuator-detector combinations, as described above. Here, at least one object can move through the effective range of an actuator of at least one of the actuator-detector combinations. This actuator can then generate a control signal based on the detection of the object. This actuator-detector combination then transmits the control signal to at least one second actuator-detector combination, which is preferably adjacent. The actuator of the second actuator-detector combination can then generate a force pulse if its detector (also) detects the presence of the object and the control signal releases the force pulse.

[0037] In a further advantageous embodiment, the method can be carried out as follows. A first actuator-detector combination can receive a control signal enabling the generation of force pulses for a specific period of time. The first actuator-detector combination can send the control signal to at least one adjacent actuator-detector combination. The control signal can then enable the generation of force pulses for the second actuator-detector combination for the said period of time. If the at least one adjacent actuator-detector combination detects the object within the period of time enabled by the control signal, the actuator of this actuator-detector combination can generate the force pulse.

[0038] In one variant, the communication signal can be used to block the neighboring actuator-detector combination. The first actuator-detector combination to detect the particle blocks the neighbors for a short time to avoid large bycatch and / or to save energy (e.g., compressed air).

[0039] The following options for conducting the procedure arise, for example: 1. Individual: Each actuator can be controlled individually, 2. Centralized cluster control: Several or all actuators are controlled via a common control unit, such as a PC. The common control unit can know the number and physical arrangement of the actuators, and in particular, neighborhood relationships, for example, can be known. 3. Decentralized cluster communication: There are two options for communication networks: 3.1 Each actuator-detector combination may have a bi-directional communication channel to its neighboring actuator-detector combination, or 3.2 Each actuator-detector combination can have a bi-directional communication channel to any other actuator-detector combination.

[0040] In case 3.1, in which each actuator-detector combination has a bidirectional communication channel to its neighbors, each actuator-detector combination has the option of transmitting a control signal or release interval transmitted to it, as well as the activation of the actuator, to its neighbors. The following scenario is therefore conceivable: Actuator-detector combination A receives a release for the period t1 to t2. Actuator-detector combination A forwards this period to its direct neighbors B and C. Within this period, an object is detected for actuator-detector combination B. Depending on the configuration, actuator-detector combination B can only use the release from actuator-detector combination A and activate its own actuator accordingly. Depending on the configuration, actuator-detector combinations B and C can, in turn, pass the information on to their neighbors.The transmitted data can contain additional information, for example how many stages you have already hiked.

[0041] In this way, additional deviations within the release period across the transport direction can be compensated.

[0042] In the case of multi-row bars of actuator-detector combinations, for example, an actuator-detector combination A can transmit a release time to the actuator-detector combination D located behind it in the transport direction. Depending on the configuration, this creates the possibility that, if the object is moving faster and the release time is too late for actuator-detector combination A, actuator-detector combination D can activate the release and, if necessary, still be able to eject the object. For objects that are too slow, this possibility also applies in reverse.

[0043] In this way, simultaneous deviations regarding the release period can also be compensated.

[0044] In case 3.2, where each actuator-detector combination has a bidirectional communication channel to every other actuator-detector combination, a multi-row arrangement of actuator-detector combinations allows the detectors used in each row to respond to different properties of the object (e.g., spectral / magnetic / (ultra-)sound). This allows multi-path sorting to be realized.

[0045] The invention can be advantageously used for sensor-supported sorting systems, such as bulk material sorting systems, in the application areas (but not limited to) food, agricultural production, recycling, minerals, and / or quality assurance in a production.

[0046] The invention will be explained below using a few examples. The features shown in the figures can also be implemented independently of the corresponding example and can be combined between the examples. The same reference numerals indicate the same or corresponding features.

[0047] It shows: Fig. 1 an example of an ejection device according to the invention, in which an upstream detector generates a control signal, Fig. 2 different examples of possible arrangements of actuator-detector combinations, and Fig. Figure 3 shows an example of an actuator-detector combination that can be used in the invention.

[0048] Fig. 1 shows a schematic sequence of a method for operating an ejection device with a plurality of actuator-detector combinations 2a, 2b, ..., 2k. The actuator-detector combinations 2a, 2b, ..., 2k each have an actuator with an effective range in which a force impulse can be exerted on an object 3 by the actuator. In addition, the actuator-detector combinations 2a, 2b, ..., 2k each have at least one detector with a detection range that at least partially overlaps the effective range of the corresponding actuator or has a distance of less than or equal to 10 mm from the effective range. In this way, the presence of the object 3 in the effective range of the corresponding actuator can be detected with the detector. In Fig. In the example shown in Figure 1, the discharge device also has an upstream sensor 4 configured to monitor objects 3 in an object stream. The upstream sensor 4 is configured to determine at least one property of the objects.

[0049] In the example shown, the actuators of the actuator-detector combination 2a, 2b, 2c ... 2k are configured to generate a force impulse directed out of the plane of the figure. The force impulse should be triggered only if, or only if, the detector of the corresponding actuator-detector pair 2a, 2b, ..., 2k detects the presence of object 3 within the effective range of this actuator. Furthermore, the upstream sensor 4 is configured to generate a control signal such that the force impulse is only released by the actuators of the actuator-detector pairs 2a, 2b, ..., 2k if a property of object 3 determined by the upstream sensor 4 matches a predetermined property.

[0050] Since a certain time elapses between the detection of the object 3 by the upstream sensor 4 until the presence of the object 3 in the effective range of one or more of the actuators, the control signal can either be generated with a delay of the corresponding period of time or contain information that informs the corresponding actuator-detector pair 2a, 2b, ..., 2k of the delay, so that the corresponding actuator is only released with a delay of the corresponding period of time.

[0051] The actuator-detector combinations 2a, 2b, ..., 2k are designed here as nozzle bars and arranged adjacent to one another along a straight line. The actuators can be compressed air nozzles. The actuator-detector pairs can be connected to one another via unidirectional or bidirectional communication channels. The straight line along which the actuator-detector combinations are arranged next to one another is perpendicular to the direction of movement of the object, i.e., perpendicular to the flow direction of the object stream. In this arrangement, the actuator-detector bar allows for fine resolution of the ejection in a direction perpendicular to the flow direction or movement direction of the object 3. At the same time, the ejection time is precisely defined due to the small extension in the direction of movement.

[0052] The process flow is shown in the upper section. In a first step S1, object 3 is detected by the upstream sensor 4. Then, in step S2, at least one property that influences the sorting decision is evaluated. Taking into account the movement of object 3 from the upstream sensor 4 to the actuator-detector bar, a rough timing and preferably also a rough position at which object 3 passes the actuator-detector bar are determined in step S3. These are then forwarded to the actuator-detector combination 2a, 2b, ..., 2k in the form of a control signal.

[0053] Fig. Figure 2 shows in the sub-figures A, B, C and D different arrays of actuator-detector combinations, which are designed here according to nozzle arrays.

[0054] Fig. Figure 2A shows a linear arrangement, i.e. a linear nozzle bar, in which the actuator-detector pairs 2a, 2b, ..., 2k are arranged adjacent to each other along a straight line perpendicular to the object flow direction.

[0055] Fig. Figure 2B shows two of the Fig. 2A, which are arranged directly one behind the other. The first bar contains nozzle-detector combinations 2aa, 2ab, ..., 2ak and the second bar contains nozzle-detector combinations 2ba, 2bb, ..., 2bk. The Fig. The design shown in Figure 2B can also be extended with additional beams.

[0056] Fig. Figure 2C shows an arrangement of actuator-detector combinations 2a, 2b, ..., 2k configured as a non-linear nozzle bar. Viewed perpendicular to the flow direction, adjacent actuator-detector combinations are arranged progressively further forward in the flow direction up to the center, and progressively further back from the center, resulting in the shape of an isosceles angle.

[0057] Fig. 2D combines the Fig. 2C with a further bar of actuator-detector pairs 2ba, 2bb, ..., 2bk arranged behind it in the flow direction, which are arranged alternately further forward and further back in the flow direction.

[0058] Fig.Figure 3 schematically shows an example of an actuator-detector combination 2 as can be used in the invention. The actuator-detector combination 2 has a detector 3b, which is embodied here, for example, as a reflex light barrier 3b. The detector 3b has a detection area 3d in which it can detect objects 3f in a bulk material flow. In the example shown, the bulk material flow runs from the top downwards, while the detection area extends from the detector 3b to the right, slightly inclined downwards, and opens at an angle.

[0059] The actuator-detector combination 2 also has an actuator 3c, which here is, for example, a compressed air nozzle 3c. The actuator 3c has an effective range 3e, within which a force impulse can be exerted on an object 3f by the actuator 3c. The effective range 3e extends horizontally to the right of the actuator, for example, and opens at an angle. The detection range 3d partially overlaps with the effective range 3e in an overlap area 3g.

[0060] In the example shown, the actuator-detector combination 2 also has a control unit 3a which is configured to control the actuator 3c in dependence on at least one signal from the detector 3b.

Claims

[1] Having a discharge device at least one actuator, wherein the actuator has an effective range in which a force impulse can be exerted on an object by the actuator when the object is located in this effective range, further comprising at least one detector, wherein the detector has a detection range which at least partially overlaps with the effective range of at least one of the actuators or has a distance of less than or equal to 10 mm from the effective range, so that the presence of the object in the effective range of this actuator can be detected with the detector, wherein the ejection device is configured to cause the actuator to generate a force impulse only when the detector detects the presence of the object in the effective range of this actuator. [2] Ejection device according to the preceding claim, wherein the ejection device is arranged to receive or generate a control signal for enabling the force impulse and is arranged to cause the actuator to generate the force impulse when the detector detects the presence of the object in the effective range of this actuator and, in addition, the control signal enables the generation of the force impulse. [3] Discharge device according to the preceding claim, which has an upstream sensor configured to observe objects in an object stream, wherein the upstream sensor is configured to determine at least one property of the objects, wherein the ejection device is configured to generate the control signal such that the force impulse of at least one of the actuators is released when the property of the object determined by the upstream sensor corresponds to a predetermined property. [4] Ejection device according to one of the preceding claims, wherein at least one of the at least one detector is configured to detect at least one property of the object in addition to the presence of the object in the effective range of the corresponding actuator, and the ejection device is configured to cause the actuator to generate the force impulse only if the detected property of the object corresponds to a predetermined property. [5] Ejection device according to one of the preceding claims, wherein at least one of the detectors is designed as a camera which is directed towards a background, wherein a colour and / or brightness of the background corresponds, within the scope of a perception accuracy of the camera, to a colour and / or brightness of objects which are not to receive a force impulse. [6] Ejection device according to one of the preceding claims, wherein the at least one actuator is a nozzle, preferably a compressed air nozzle and / or a water jet nozzle and / or at least one mechanical finger and / or at least one mechanical flap. [7] Ejection device according to one of the preceding claims, wherein the at least one detector comprises or is an optical sensor, a light barrier, a reflex light barrier, a proximity switch, a capacitive sensor, an eddy current sensor and / or a magnetic sensor. [8] Ejection device according to one of the preceding claims, wherein the effective ranges of the actuators are each observed by a separate one of the detectors and / or the detection ranges of the detectors each overlap at least partially with exactly one effective range of one of the actuators. [9] Ejection device according to one of the preceding claims, comprising a plurality of actuator-detector combinations, wherein each of the actuator-detector combinations contains one of the actuators and one of the detectors, the detection range of which overlaps at least partially with the effective range of the actuator of the same actuator-detector combination. [10] Ejection device according to the preceding claim, comprising a plurality of actuator-detector combinations which are arranged next to one another along a straight line, preferably forming an actuator bar, or which are arranged next to one another along a plurality of parallel straight lines, preferably forming a plurality of parallel actuator bars. [11] Ejection device according to one of claims 9 or 10, wherein some or all of the actuator-detector combinations have a uni- or bi-directional communication channel with at least one adjacent one of the actuator-detector combinations, preferably with all immediately adjacent actuator-detector combinations, wherein control signals can be transmitted between the actuator-detector combinations via the communication channels, wherein at least some or all of the actuator-detector combinations are arranged to cause their actuator to generate a force impulse only when the control signal enables the generation of the force impulse. [12] Ejection device according to one of the two preceding claims, comprising at least two parallel actuator bars, each having a plurality of actuator-detector combinations, wherein an actuator-detector combination of a first of the at least two actuator bars and an actuator-detector combination of a second of the at least two actuator bars are arranged one behind the other in a flow direction in which the objects move, wherein the actuator-detector combination which is at the rear in the flow direction is arranged to receive a control signal from the actuator-detector combination which is at the front in the flow direction, wherein the actuator-detector combination located at the rear in the direction of flow is configured to cause the actuator of this actuator-detector combination to generate the force impulse when the detector of this actuator-detector combination detects the presence of an object in the effective range of this actuator and, at the same time, the control signal enables the generation of the force impulse. [13] Ejection device according to one of the preceding claims, comprising a control unit which is configured to control the at least one actuator in dependence on at least one signal of at least one of the detectors and / or a control signal. [14] Method for operating a discharge device according to one of the preceding claims, wherein at least one object moves through the effective range of at least one of the actuators, the presence of the object in the effective range of the actuator is detected by at least one of the detectors and the actuator generates a force impulse at most when the detector detects the presence of the object in the effective range of this actuator. [15] Method according to the preceding claim, wherein the ejection device is an ejection device according to claims 9 to 14, wherein at least one object moves through the effective range of an actuator of at least one of the actuator-detector combinations, this generates a control signal based on the detection of the object, this actuator-detector combination transmits the control signal to at least a second of the actuator-detector combinations, preferably an adjacent one, the actuator of the second actuator-detector combination generates a force impulse when its detector detects the presence of the object and the control signal releases the force impulse. [16] Method according to one of the two preceding claims, wherein a first actuator-detector combination receives a control signal enabling the generation of force pulses for a specific period of time, wherein the first actuator-detector combination sends a control signal to at least one adjacent actuator-detector combination, with which the specific period of time for generating force pulses by the actuator of the adjacent actuator-detector combination is released, wherein, if the at least one adjacent actuator-detector combination detects an object in the period enabled by the control signal, the actuator of this adjacent actuator-detector combination generates a force impulse.

Images