Photoelectric sensor for detecting objects in monitoring area

Abstract

The invention relates to a photosensor, in particular a time-of-flight camera or LiDAR sensor, for detecting at least one object within a monitoring area, the photosensor comprising a light emitter, a light receiver and an evaluation unit. The light emitter is configured to emit transmitted light to a monitoring area. The light receiver is configured to receive the received light retroreflected by the monitoring area. An evaluation unit configured to obtain distance data about the monitored area based on the received light, where the distance data comprises an intensity value and an associated distance value; for all data portions of the distance data not originating from the object, in particular data portions having an intensity value less than a first intensity limit value, the distance to the object, in particular the distance to the edge of the object, is determined in each case on the basis of the distance data, and the distance to the edge of the object is determined in each case on the basis of the determined distance of the respective data portion to the object. Based on a scattered light behavior model of the photoelectric sensor and based on the intensity value of the corresponding data part, determining a probability value that the corresponding data part represents interference; and identifying, as interference, all the data portions having the determined probability values equal to or greater than the probability limit value.

Description

Technical Field

[0001] The present invention relates to a photoelectric sensor, particularly a time-of-flight camera or lidar sensor, for detecting at least one object within a monitoring area, wherein the photoelectric sensor includes means for identifying interference in a portion of the acquired distance data about the monitoring area. Background Technology

[0002] Photoelectric sensors can be used in industrial safety applications to achieve environmental perception of a monitored area, particularly a three-dimensional environmental perception, thereby improving the safety and efficiency of industrial processes in industrial plants. Examples of such photoelectric sensors include Time-of-Flight (ToF) cameras and LiDAR (Light Detection and Ranging) sensors. For example, photoelectric sensors can be fixedly attached to an industrial plant or to robots that can move autonomously within the plant. Industrial plants can be production workshops, warehouses, power plants, chemical plants, food processing plants, or livestock facilities. Reflectors, especially retroreflectors, can be attached to industrial plants and used for obstacle marking / highlighting and / or navigation functions. They can also be used by photoelectric sensors on autonomous mobile robots (e.g., AGVs – autonomous guided vehicles) to control, locate, and / or navigate the robots.

[0003] If high-reflectivity (especially reflective) objects, such as reflectors (particularly retroreflectors), warning vests, metal, or reflective objects, are present in the monitoring area, the received light in the typical receiving optics of a photoelectric sensor may be partially scattered at the lens edges and / or other optical elements. This is particularly true for photoelectric sensors that scan the environment simultaneously rather than making separate measurements. Scattered light appears as image interference around objects (especially against a background with very weak or distant reflection) and can cause distortion in distance measurements. Image interference may trigger warning or protection fields configured in the photoelectric sensor in an undesirable manner, thus causing the autonomous robot to stop unnecessarily. This reduces the robot's usability for its intended use and / or may even render the photoelectric sensor completely unusable in industrial plants, as it repeatedly triggers the robot's safety stop at the same location in the industrial plant (i.e., near the reflector).

[0004] Known photoelectric sensors attempt to overcome or avoid image interference by using receiving lenses with less internal reflection. However, such receiving lenses are complex, expensive, difficult to implement, and / or, even if possible, only prevent image interference to a certain extent. Other known photoelectric sensors reduce the intensity of transmitted light. However, reduced transmitted light typically leads to reduced distance, reduced field of view, detection loss, and / or accuracy loss. Summary of the Invention

[0005] This invention is based on the aim of providing an improved photoelectric sensor, particularly for identifying image interference.

[0006] A photoelectric sensor having the features of claim 1 is provided to achieve this objective.

[0007] According to the invention, a photoelectric sensor, particularly a time-of-flight camera or lidar sensor, for detecting at least one object within a monitoring area (especially for vehicle navigation), includes a light emitter, a light receiver, and an evaluation unit. The light emitter is configured to emit transmitted light into the monitoring area. The light receiver is configured to receive the received light reflected back from the monitoring area. The evaluation unit is configured to obtain distance data about the monitoring area based on the received light, wherein the distance data includes intensity values ​​and associated distance values. The evaluation unit is further configured to: for all data portions (e.g., pixels) in the distance data that are not derived from an object, particularly for all data portions in the distance data whose intensity values ​​are less than a (predetermined) first intensity limit value, in each case, determine the distance to the object, particularly the distance to the edge of the object, based on the distance data; and in each case, based on the determined distance to the object of the corresponding data portion, based on a scattered light behavior model of the photoelectric sensor, and based on the intensity value of the corresponding data portion, determine a probability value that the corresponding data portion represents interference. The evaluation unit is further configured to identify all data portions with determined probability values ​​equal to or greater than the (predetermined) probability limit value as interference.

[0008] In other words, this invention is based on the understanding that for all data portions in distance data that do not originate from (strong) reflective objects (e.g., reflectors, especially retroreflectors), in each case, the probability value of the corresponding data portion representing interference can be determined based on the distance between the data portion and the object (especially the distance to the edge of the object), the intensity value of the data portion, and a calibrated or simulated scattered light model. Based on these determined probability values ​​(which can also be referred to as a probability mask or probability map in the overall context of all data portions), data portions in the distance data that have a certain probability of representing interference can be identified and filtered out.

[0009] Data portions in distance data that do not originate from objects can be identified using common image segmentation algorithms, particularly machine learning algorithms. Specifically, the fact that interference caused by scattered light typically has a lower intensity than the reflected object itself can be utilized. Additionally or alternatively, data portions in distance data that do not originate from objects can therefore be identified (and segmented) by filtering using a (first) intensity limit value. For all data portions in the distance data with intensity values ​​below the (first) intensity limit value (i.e., data portions not originating from objects), a scattered light model can be used (individually) to determine the probability value that the corresponding data portion represents interference. Based on these determined probability values, data portions representing interference can be filtered out.

[0010] A challenge in filtering distance data lies in the difficulty of identifying whether the measurement data is valid (e.g., whether it originates solely from a dark scene) or whether it actually represents interference. This is particularly relevant for safety-related applications, as all measurement data that cannot be reliably identified as erroneous or interfering must be retained for functional safety reasons. In this regard, the closer a pixel in the distance data is to a strongly reflective object, the less efficient traditional filters tend to be. Furthermore, the larger the strongly reflective object in the monitoring area, the more pronounced the interference may be. The photoelectric sensor based on probability limit filtering according to the present invention can be more efficient and / or more reliable in identifying image interference, especially when strongly reflective objects are present in the monitoring area.

[0011] A model of the scattered light behavior of a photoelectric sensor can be calibrated, created, or derived by means of calibration with actual measurement data. The scattered light behavior of a photoelectric sensor can be understood through its interaction with objects (e.g., reflectors) within the monitoring area. In other words, it can be specified that only certain strongly retroreflective objects (especially retroreflectors) are located within the monitoring area. By recording distance data about the monitoring area with and without objects, a model of the scattered light behavior can be calibrated, created, or derived. Additionally or alternatively, the model can also be calibrated, created, or derived using simulation. It is also conceivable that the model is calibrated, created, or derived using machine learning algorithms. Using a scattered light behavior model calibrated, created, or derived in this way, the expected intensity deviation for each data segment (e.g., each pixel) can be calculated (e.g., based on the object position and size in the distance data). If a model of the scattered light behavior of a photoelectric sensor is calibrated, created, or derived, the model can also be transferred to other photoelectric sensors with the same or similar receiving optics. The calibration process can be automated using a device that repositions objects within the monitoring area.

[0012] The distance values ​​in the distance data portion can correspond to the distance measured in meters between the monitoring area (or objects within the monitoring area) and the photoelectric sensor. The distance between the data portion and the object (or the edge of the object) can correspond to the distance measured in pixels (on the sensor), or it can be converted to a distance in meters.

[0013] The first intensity limit value can be, for example, a value in any unit, such as 20,000 AU or 20,000 counts. Then, the data portion of the object distance data can have an intensity value equal to or greater than the first intensity limit value. The first intensity limit value can correspond to the maximum intensity value that the photoelectric sensor can measure. In this case, the intensity value in the object distance data can take the value of the first intensity limit value and can be referred to as "overmodulation".

[0014] The first intensity limit value can correspond to a percentage of the maximum intensity that the photoelectric sensor can measure, for example, it can correspond to 50% of the maximum measurable intensity, preferably 60%, preferably 70%, preferably 80%, preferably 90%, preferably 95%, preferably 99%, or even more preferably 100% (i.e., saturation value).

[0015] The probability limit value can be, for example, 50% (or 0.5), preferably 60%, preferably 70%, preferably 80%, preferably 90%, or even more preferably 95%. The probability limit value can preferably be equal to or greater than 0.5 and less than or equal to 0.7.

[0016] It is understood that the evaluation unit can be part of a photoelectric sensor or an external computing unit, such as a server, and the photoelectric sensor communicates with the external computing unit via a (wireless or wired) signal connection.

[0017] Photoelectric sensors are preferably safety sensors, specifically safety ToF cameras or safety LiDAR sensors. In this context, the term "safety" or "security" can be understood within the specific safety standard meanings of standards such as DIN ISO 13849 or DIN EN 61508. Therefore, photoelectric sensors can control errors within a certain safe level.

[0018] The transmitted light is reflected back by objects (and other items) within the monitored area and used as received light.

[0019] According to one embodiment, the object includes a reflector (e.g., a retroreflector). Specifically, the object can be a reflector, and preferably a retroreflector. Regardless of the angle of incidence, the retroreflector preferably reflects the transmitted light back to the emitter. The retroreflector may include multiple small-angle mirrors or hemispherical mirrors that reflect light back to the direction of the light emitter, preferably with only a few degrees of small scattering. The retroreflectivity of the reflector can be equal to or greater than 80%, preferably equal to or greater than 85%, preferably equal to or greater than 90%, preferably equal to or greater than 95%, preferably equal to or greater than 99%. The intensity values ​​of the reflectors in the distance data can have a maximum measurable intensity value (regardless of the distance between the reflector and the photoelectric sensor), i.e., they can (always) be overmodulated. The reflectors (especially retroreflectors) can be installed in industrial plants for controlling, locating, and / or navigating autonomous robots. It can be specified that only certain reflectors (especially retroreflectors) are located within the monitoring area.

[0020] The evaluation unit can be configured to determine a refraction value based on distance data, where the refraction value can be estimated based on the obtained intensity value and distance value, or it can be calculated based on the obtained intensity value and distance value. For example, the corresponding refraction value can be proportional to the square of the obtained intensity value multiplied by the obtained distance value (in meters). Generally, retroreflectance can be estimated based on the measured intensity and distance of the object to determine the "reflectivity of the observed surface." In other words, it is an estimate of the material properties of the observed surface (or object) identified by the photoelectric sensor. Furthermore, the retroreflectance value can be normalized based on the intensity of the transmitted light.

[0021] According to one embodiment, the distance data includes multiple pixels, each pixel having an intensity value and an associated distance value.

[0022] According to one embodiment, the evaluation unit is configured to determine the distance, size, retroreflectivity, and / or intensity of the reflected object based on distance data. The distance to the object may refer to the distance between the object and the photoelectric sensor.

[0023] According to one embodiment, the probability value P representing interference in the corresponding data portion is determined using the following Equation 1. BP ,

[0024] [Equation 1]

[0025] Among them, I BP This refers to the measured intensity value of the corresponding data section, x. BP It is the distance (in pixels) between the corresponding data portion and the object (especially the edge of the object), and F(x) BP ) is based on x BPThe scattered light value is determined using the scattered light behavior model F. In other words, for example, I BP This can correspond to the basic intensity value I originating from a monitoring area without objects. NoObj and the intensity value I derived from the object Obj The sum of, where F(x) BP This can correspond to the intensity value I from the object. Obj Probability value P BP It can also depend on or correspond to 1 minus I NoObj with I BP The ratio (i.e., 1 - I) NoObj / IBP Even if I don't necessarily know NoObj It is still possible to know I NoObj with I BP The ratio will vary depending on the object. For example, if I NoObj with I BP If the objects are very similar or completely identical (i.e., no objects are present in the monitored area), then P BP The value is close to zero. If the basic strength value I NoObj If P is close to zero, then BP The value is close to 1. Therefore, the closer the data portion is to the object, the more likely it is to represent interference. A higher basic intensity value I... NoObj This will result in a lower probability of interference and / or interference from objects in the corresponding data.

[0026] According to one embodiment, the evaluation unit is configured to determine a probability value for a corresponding pixel to represent interference, the probability value depending on the determined distance between the corresponding data portion and the object (particularly the edge of the object), the intensity value of the corresponding data portion, the maximum measurable intensity (e.g., at a distance where the data portion is zero from the object and / or at a distance where the object is zero from the photoelectric sensor), and the size of the object.

[0027] According to one embodiment, the evaluation unit is configured to determine the size of the retroreflected object based on the portion of the distance data in which the intensity value is equal to or greater than a first intensity limit value.

[0028] According to one embodiment, the evaluation unit is configured to identify the data portion representing interference in the distance data only if the size of the determined object is equal to or greater than a predetermined size limit, the retroreflectivity of the determined object is equal to or greater than a predetermined intensity limit, and / or the intensity of the determined object is equal to or greater than a second intensity limit.

[0029] The retroreflectance limits described herein can be measured and / or determined using test samples based on specific safety standards. According to specific safety standards, for example, it may be required that an object can be reliably identified even if its retroreflectance is 4%. Therefore, for example, the retroreflectance limit can be set to 4% because objects with low retroreflectance can be ignored.

[0030] The second intensity limit value can be equal to, higher than, or lower than the first intensity limit value. For example, the size of the object can be measured based on the number of pixels, or converted based on the (real) spatial range. The size of the object (e.g., a retroreflector) is known. Furthermore, it is known that the object may be a very bright object, i.e., a strongly reflective object. Based on this, it is possible to identify whether such an object with specific properties (e.g., a retroreflector) is located within the monitoring area. If so, interference identification or identification (and filtering) of the data portion representing interference is performed. Otherwise, the evaluation unit can cancel the evaluation of the distance data at this time, can skip interference identification, and / or can emit a signal. This selective identification (and filtering) procedure can be particularly important for security applications, as it is essential to avoid unintentionally filtering out valid measurement data (i.e., the data portion of a real object) as much as possible. In addition, this approach can improve the efficiency of photoelectric sensors. If multiple sufficiently large and sufficiently reflective (bright) objects (e.g., retroreflectors) are identified, the interference identification process described herein can be iteratively implemented for each of these objects.

[0031] According to one embodiment, the evaluation unit is configured to determine the size of an object based on pixels in distance data whose intensity values ​​are equal to or greater than a first intensity limit value, wherein the size of the object is preferably determined as the number of pixels. The evaluation unit is configured to: for each pixel whose intensity value is less than the first intensity limit value, determine the (minimum) distance to the nearest pixel of the object, wherein the distance to the object is preferably determined as the number of pixels. The evaluation unit is configured to determine a probability value that a corresponding pixel represents interference based on the intensity value of the corresponding pixel, the determined distance between the corresponding pixel and the nearest pixel of the object, the maximum measurable intensity and / or the intensity of the object, and the size of the object, and to identify all pixels whose probability values ​​are equal to or greater than the probability limit value as interference.

[0032] According to one embodiment, the evaluation unit is configured to determine the distance of the retroreflecting object (to the photoelectric sensor) based on distance data, and to identify all data portions with probability values ​​equal to or greater than a probability limit value, and whose associated distance values ​​lie within a tolerance range around the determined object's distance, as interference. This procedure is based on the understanding that data portions representing interference typically appear at approximately the same location as the (real) retroreflecting object's distance to the photoelectric sensor. Because additional filtering is performed based on the tolerance range around the determined object's distance (or in other words, based on a distance corridor), the filtering results can be rationalized according to the probability limit value. If multiple retroreflecting objects exist within the monitoring area, the process can be performed and / or iterated separately for each identified object.

[0033] The distance to an object can refer to its (average) relative distance to the photoelectric sensor. For example, the tolerance range can include a range of ±50 cm, preferably ±20 cm, preferably ±12 cm, preferably ±10 cm, and preferably ±5 cm around the determined distance to the object.

[0034] According to one embodiment, the tolerance range includes ±20%, preferably ±15%, preferably ±10%, preferably ±6%, preferably ±5%, and preferably ±1% of the value of the distance between the determined object and the determined object (with the photoelectric sensor).

[0035] According to one embodiment, the evaluation unit is configured to input each pixel in the distance data whose intensity value is equal to or greater than a first intensity limit value into a distance histogram, identify peaks in the distance histogram, wherein the peaks are preferably the largest peaks (i.e. vertices) in the distance histogram, determine the size of the object based on the number of pixels below the peaks, and determine the distance of the object based on the position of the peaks in the distance histogram.

[0036] In other words, the distance data is searched for pixels with intensity values ​​equal to or greater than a first intensity limit. Such pixels can then be referred to as overmodulated pixels. Pixels affected by (retro)reflectors are typically or even frequently overmodulated. Based on the identified overmodulated pixels, objects with strong retroreflection (such as retroreflectors) can then be identified, especially.

[0037] According to one embodiment, data portions identified as interference in the distance data (particularly used to control the robot's movement) are removed from the distance data, marked as invalid, and / or ignored during further evaluation of the distance data.

[0038] Another subject of the invention is a system comprising at least one photoelectric sensor as described herein and at least one autonomous robot, wherein the photoelectric sensor is preferably attached to and movable with the autonomous robot.

[0039] According to one embodiment, the system includes a reflector, preferably a retroreflector. The reflector (especially a retroreflector) can be attached to a monitoring area, where a robot can move. The reflector (especially a retroreflector) can be used as an obstacle marking / obstacle highlighting and / or navigation function, and can be used by photoelectric sensors for controlling, locating, and / or navigating the robot.

[0040] Another subject of the invention is the use of the photoelectric sensor described herein for detecting at least one object in a monitored area.

[0041] Another aspect of the invention is a method for detecting at least one object in a monitoring area, wherein transmitted light is emitted into the monitoring area; wherein received light reflected back from the monitoring area is received; wherein distance data about the monitoring area is obtained based on the received light, and in particular, the distance data is measured by means of a time-of-flight method, wherein the distance data includes an intensity value and an associated distance value. Furthermore, for all data portions in the distance data that do not originate from an object (and in particular for all data portions whose intensity values ​​are less than a first intensity limit), in each case, the distance to the object, in particular the distance to the edge of the object, is determined based on the distance data; and in each case, based on the determined distance of the corresponding data portion to the object, a scattered light behavior model of a photoelectric sensor, and the intensity value of the corresponding data portion, a probability value indicating that the corresponding data portion represents interference is determined. Then, all data portions in the distance data whose determined probability values ​​are equal to or greater than a predetermined probability limit value are identified as interference.

[0042] According to one embodiment, the object includes a reflector, preferably a retroreflector.

[0043] According to one embodiment, distance, retroreflectivity, intensity, and / or the size of the retroreflected object are determined based on distance data.

[0044] According to one embodiment, the probability value that the corresponding data portion represents interference is determined based on the determined distance between the corresponding data portion and the object, the intensity value of the corresponding data portion, the maximum measurable intensity (e.g., when the distance between the data portion and the object is equal to zero and / or when the distance between the object and the photoelectric sensor is equal to zero), and the size of the object.

[0045] According to one embodiment, all data portions whose probability values ​​are equal to or greater than a probability limit and whose associated distance values ​​are within the tolerance range of the determined object's distance are identified as interference.

[0046] It is understood that the description of the photoelectric sensor according to the present invention also applies to the use, systems, and methods of the photoelectric sensor. This is particularly true of the embodiments and advantages. Furthermore, it is understood that, unless otherwise expressly stated, all features and embodiments disclosed herein may be combined. Attached Figure Description

[0047] The invention will now be described by way of example only, with reference to possible embodiments and the accompanying drawings. As shown in the figures:

[0048] Figure 1 This is a schematic diagram of a photoelectric sensor according to an embodiment of the present invention;

[0049] Figure 2A This is a bird's-eye view representation of distance data about the monitoring area obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0050] Figure 2B This is a bird's-eye view representation of distance data about a monitoring area (with objects that emit strong backlight) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0051] Figure 3A This is a center perspective representation of distance data about a first monitoring area (with a first test object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0052] Figure 3B This is a center perspective representation of distance data about a second monitoring area (with a second test object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0053] Figure 3C This is a center perspective representation of distance data about a third monitoring area (with a third test object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0054] Figure 4A This is a center perspective representation of distance data about a first monitoring area (without a first test object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0055] Figure 4B This is a center perspective representation of distance data about a second monitoring area (without a second test object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0056] Figure 4C This is a center perspective representation of distance data about a third monitoring area (without a third test object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0057] Figure 5AThe center perspective representation of the distance data for the first test object, which has been adjusted according to the basic intensity;

[0058] Figure 5B The center perspective representation of the distance data for the second test object, which has been adjusted according to the basic intensity;

[0059] Figure 5C The center perspective representation of the distance data for the third test object, which has been adjusted according to the basic intensity;

[0060] Figure 6A This is a representation of a portion of the distance data for the first test object, which has been adjusted based on the base intensity;

[0061] Figure 6B This is a representation of a portion of the distance data for the second test object, which has been adjusted based on the base intensity;

[0062] Figure 6C This is a representation of a portion of the distance data for the third test object, which has been adjusted based on the base intensity.

[0063] Figure 7A This represents the average portion of the distance data of the first test object along the y-axis.

[0064] Figure 7B This represents the average portion of the distance data of the second test object along the y-axis.

[0065] Figure 7C This represents the average portion of the distance data of the third test object along the y-axis.

[0066] Figure 8A A graphical representation of a light scattering behavior model of a photoelectric sensor according to an embodiment of the present invention, the model being calibrated relative to a first test object;

[0067] Figure 8B A graphical representation of a light scattering behavior model of a photoelectric sensor according to an embodiment of the present invention, the model being calibrated relative to a second test object;

[0068] Figure 8C A graphical representation of a light scattering behavior model of a photoelectric sensor according to an embodiment of the present invention, the model being calibrated relative to a third test object;

[0069] Figure 9A This is a center perspective representation of the distance values ​​of the monitoring area (where there is an object) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0070] Figure 9BThis is a central perspective representation of the intensity values ​​of distance data of an object obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0071] Figure 9C A representation of the segmentation of distance data;

[0072] Figure 9D Center perspective representation of the distance between the segmented pixels and the object;

[0073] Figure 10A To represent the scattered light for modeling objects;

[0074] Figure 10B The central perspective representation of the determined probability diagram;

[0075] Figure 10C Representation of the identified image interference;

[0076] Figure 10D The center perspective representation of the distance values ​​for the object's distance data has been adjusted to account for image interference;

[0077] Figure 11A This is a bird's-eye view representation of distance data about a monitoring area (without objects) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0078] Figure 11B This is a center perspective representation of distance data about a monitoring area (with objects) obtained by means of a photoelectric sensor according to an embodiment of the present invention;

[0079] Figure 11C This is a bird's-eye view representation of distance data about a monitored area (with objects) obtained by means of a photoelectric sensor according to an embodiment of the present invention. Detailed Implementation

[0080] Figure 1A photoelectric sensor 100, particularly a time-of-flight camera or LiDAR sensor, is shown, according to an embodiment, for detecting at least one object 40 within a monitoring area. The photoelectric sensor 100 includes a light emitter 10, a light receiver 20, and an evaluation unit 30. The light emitter 10 is configured to emit transmitted light 11 into the monitoring area. The light receiver 20 is configured to receive received light 12 reflected from the monitoring area, and particularly received light reflected from the object 40 within the monitoring area. The evaluation unit 30 is configured to obtain distance data about the monitoring area based on the received light 12, wherein the distance data includes an intensity value and an associated distance value. The evaluation unit 30 is also configured to: determine, for all data portions of the distance data that do not originate from the object 40 (e.g., pixels), and in particular for all data portions of the distance data whose intensity values ​​are less than a first intensity limit value, the corresponding distance to the object 40 (e.g., in pixels), especially the distance to the edge of the object 40; and in each case, based on the determined distance to the corresponding data portion, based on the scattered light behavior model of the photoelectric sensor 100, and based on the intensity value of the corresponding data portion, determine the probability value that the corresponding data portion represents interference; and identify all data portions whose determined probability values ​​are equal to or greater than the probability limit value as interference.

[0081] like Figure 1 As shown, the evaluation unit 30 of the photoelectric sensor can be configured to determine the distance (distance to the photoelectric sensor, e.g., in meters), size, intensity, and / or retroreflectivity of the retroreflecting object 40 based on distance data. For example, the evaluation unit 30 can determine the size of the retroreflecting object 40 based on the portion of the distance data where the intensity value is equal to or greater than a first intensity limit value.

[0082] Figure 1 The evaluation unit 30 of the photoelectric sensor 100 shown can be configured to determine the distance of the object 40 and / or identify the data portion representing interference in the distance data only if the size of the object 40 is equal to or greater than a predetermined size limit, the intensity of the object 40 is equal to or greater than a predetermined second intensity limit, and / or the retroreflectance of the object is greater than or equal to a retroreflectance limit. In this way, the effective data portion (i.e., the data portion of the actual object) can be avoided from being unintentionally filtered out, which is particularly important for security applications. If multiple objects that are sufficiently large and have sufficient retroreflectance are identified or detected, the interference identification process can be iteratively performed on each of the identified objects.

[0083] For example, Figure 1As shown, the evaluation unit 30 of the photoelectric sensor 100 can be configured to identify all data portions (e.g., pixels) whose probability values ​​are equal to or greater than the probability limit value and whose associated distance values ​​are within the tolerance range around the determined object 40 (with the photoelectric sensor) as interference.

[0084] Figure 2A and Figure 2B A bird's-eye view representation of distance data about the monitored area (with or without strongly reflective objects 40) obtained by means of the photoelectric sensor 100 is shown, such as Figure 1 As shown. Figure 2A and Figure 2B As shown, the obtained distance data can include multiple pixels, each with an intensity value and an associated distance value.

[0085] Especially when the monitoring area is scanned simultaneously (as opposed to scanning in the case of individual measurements), scattered light from the optics of the light receiver 20 of the photoelectric sensor 100 can make accurate distance measurements more difficult or even prevent them. A strong retroreflection rate of the reflected object 40 (e.g., a retroreflector) to the transmitted light can cause distorted distance measurements around the image area of ​​the object 40. This image interference 21 typically occurs when the energy of the scattered light is higher than the baseline level or intensity reflected back from the monitoring area, especially against dark or distant backgrounds.

[0086] Figure 2A and Figure 2B The scene of a chair in front of multiple windows illustrates this point, with the windows serving as a substitute for a dark background. For example... Figure 2A As shown, noise in the distance values ​​can be observed in the window area. However, this noise is still easily identifiable and therefore controllable. Figure 2B As shown, when a strongly retroreflective object 40 (in this example, a retroreflector) is placed on the chair, all noise values ​​or image interference 21 are assumed to be at the distance of object 40. Scattered light will cover some or all of the pixels in the environment of object 40 that do not have sufficient self-luminous intensity. This produces spherical image interference or artifacts formed by pixel clouds 21, where the radius of the spherical artifact can correspond to the distance of object 40, and the center can correspond to the position of photoelectric sensor 100. Figure 2A and 2B (Not shown in the image).

[0087] in short, Figure 2BThe pixel 21 shown may be generated due to multiple reflections and scatterings of received light by optical elements in the light receiver 20 of the photoelectric sensor 100. These pixels 21 represent interference and can trigger warning or protection fields configured in the photoelectric sensor 100. Therefore, this could unnecessarily cause the safe stopping of an autonomous robot that can move within an industrial plant, particularly within a monitored area. It is understood that... Figure 1 As shown, the photoelectric sensor 100 can be attached to an autonomous robot and, in particular, can move with it.

[0088] The procedure described herein allows for the identification (and filtering) of pixels covered by scattered light with a certain probability. In this regard, a model of the behavior of scattered light from the photoelectric sensor 100 is used, which can also be viewed as an approximation of the area affected by the interaction between the retroreflecting object 40 and the optics of the light receiver 20.

[0089] Figures 3A to 8C This document explains how to create and / or calibrate a model of the scattered light behavior of a photoelectric sensor for a specific object in a monitoring area. The procedure described herein is based on the fact that the scattered light behavior of the optical elements of the photodetector 20 of the photoelectric sensor 100 (e.g., ...) Figure 1 (As shown) can be understood through interaction with certain (test) objects 41, 42, and 43. Based on distance measurements of the same monitoring area with and without test objects, the scattered light behavior models 41, 42, and 43 can be calibrated for specific test objects. The test object described herein can be a retroreflector. Calibration can be performed by recording actual measurement data, such as... Figures 3A to 8C As illustrated. Additionally or alternatively, a model of scattered light behavior can also be simulated.

[0090] Figure 3A , Figure 3B and Figure 3C The diagram shows a center perspective representation, or more precisely, intensity values, of distance data for the first, second, and third monitoring areas (where strong backscattering occurs in the first test object 41, second test object 42, and third test object 43). This data was obtained using the photoelectric sensor 100, such as… Figure 1 As shown. The evaluation unit 30 of the photoelectric sensor 100 determines the size of the first test object 41 as 7971 pixels, the size of the second test object 42 as 1055 pixels, and the size of the third test object 43 as 2964 pixels.

[0091] Figure 4A , Figure 4B and Figure 4CThe diagram shows a center perspective representation of the distance data for the first, second, and third monitoring areas (where no test objects 41, 42, and 43 are present), or more precisely, the intensity values ​​in the distance data, obtained using the photoelectric sensor 100, such as... Figure 1 As shown. Figure 4A , Figure 4B and Figure 4C The distance data shown represents a record of the basic level or basic intensity of the corresponding monitoring area.

[0092] Figure 5A , Figure 5B and Figure 5C A central perspective representation of the distance data (or more precisely, the intensity values) of the first test object 41, the second test object 42, and the third test object 43 is shown, respectively, the distance data having been adjusted (e.g., by subtraction) according to the corresponding basic intensity. Figure 5A , Figure 5B and Figure 5C In the text, selected portions 31, 32, and 33 are each marked with a rectangle, where these portions are within... Figure 6A , Figure 6B and Figure 6C The following is shown in more detail. Sections 31, 32, and 33 show the intensity development from the edge of the corresponding test object 41, 42, and 43 to the outside, particularly to the corresponding monitoring area (left side).

[0093] like Figure 7A , Figure 7B and Figure 7C As shown, the average value of selected portions 31, 32, and 33 is taken along the y-axis. Data points 51, 52, and 53 obtained in this way are... Figure 8A , Figure 8B and Figure 8C The data is graphically marked with a cross shape. To approximate the intensity deviation caused by scattered light, the fitting functions 61, 62, and 63 can be approximated to the data points in each case according to the following general equation 2.

[0094] [Equation 2]

[0095] .

[0096] In Equation 2, x represents the distance (e.g., measured in pixels) between the corresponding data point and the test objects 41, 42, and 43, specifically the distance to the edges of the test objects 41, 42, and 43. Parameter a can take values ​​between -2.3 and 2.6, parameter b can take values ​​between 0.2 and 0.3, and parameter c is calculated according to Equation 3 based on the maximum measurable intensity / max of the photoelectric sensor 100 (here, 20000 AU).

[0097] [Equation 3]

[0098] .

[0099] Equation 2 can also be regarded as a general model of the scattered light behavior of the photoelectric sensor 100.

[0100] exist Figure 8A , Figure 8B or Figure 8C In this context, the first fitting function 61 of the first test object 41, the second fitting function 62 of the second test object 42, and the third fitting function 63 of the third test object 43 are each marked as continuous lines. Figure 8A In the first fitting function 61 of the first test object 41, shown graphically, the value of parameter a is -2.3432, the value of parameter b is 0.21166, and the value of parameter c is 9.9035. Figure 8B In the second fitting function 62 of the second test object 42, which is shown graphically, the value of parameter a is -2.4046, the value of parameter b is 0.27182, and the value of parameter c is 9.9035. Figure 8C In the third fitting function 63 of the third test object 43, which is shown graphically, the value of parameter a is -2.5399, the value of parameter b is 0.22513, and the value of parameter c is 9.9035. In this way, the scattered light behavior model of the photoelectric sensor 100 can be calibrated for each known (test) object, such as for each known reflector type.

[0101] The measurement results show that the values ​​of parameters a and b in the scattered light behavior model according to Equation 2 can depend on the size (i.e., number of pixels) of the test objects 41, 42, and 43. In other words, parameter a can be selected based on the size of the test objects (or the determined number of pixels of the test objects) and can take a value between -2.3 and -2.6. Additionally or alternatively, parameter b can be selected based on the size of the test objects (or the determined number of pixels of the test objects) and can take a value between 0.2 and 0.3. The smaller the test objects 41, 42, and 43, the larger the value of parameter b can be. This means that for other objects that have not yet been calibrated, the values ​​of parameters a and b can be selected based on the determined size of object 40, for example, by using a pre-calibrated function or interpolation of values ​​taken from a pre-created lookup table.

[0102] Figures 9A to 11C This explains how to identify and filter interference in the distance data obtained by the photoelectric sensor 100.

[0103] Figure 9A and Figure 9BThe image shows a central perspective representation of distance or intensity values ​​in distance data about a monitored area (with object 40) obtained by means of a photoelectric sensor 100, such as... Figure 1 As shown.

[0104] For example, it is possible to identify and segment (and / or binarize) data that is not derived from intensity limit values. Figure 9A or Figure 9B The pixels of object 40 shown. For example, all pixels with intensity values ​​less than 90%, preferably 95%, preferably 99%, or even more preferably 100% of the maximum intensity that the photoelectric sensor can measure (20000 AU in this example) can be identified, and as Figure 9C As shown in black in the image, pixels can be set to a value of zero during segmentation. The remaining pixels, whose intensity value is equal to or greater than the intensity limit (i.e., pixels with an object intensity of 40), are assigned a value of 1 during segmentation. Figure 9C In the image, the 40 pixels of the object are shown as white, with an intensity limit of 19000 AU applied.

[0105] For all segmented pixels (i.e., pixels not originating from object 40), determine the (minimum) distance to the nearest pixel of object 40, such as... Figure 9D As shown. It is well known that object 40 can correspond to the first test object 41, therefore the scattered light behavior model (i.e., from...) Figure 8A The fitting function 61 has been calibrated relative to object 40. Figure 10A The scattered light relative to object 40 is shown, and this scattered light is based on a known model of scattered light behavior (i.e., from...). Figure 8A The approximation is obtained using the fitting function 61). For each segmented pixel, the probability value that the corresponding pixel represents image interference can be determined based on its determined distance from object 40, its intensity value, and the scattered light approximated according to the pixel distance model. More precisely, the approximate scattered light at the distance of the corresponding pixel is divided by the intensity value of the corresponding pixel (according to Equation 1). This procedure is based on the understanding that the larger the original intensity value of the scattered light of a pixel relative to object 40, the less likely that pixel is covered by scattered light. Conversely, pixels with low intensity values ​​near object 40 are more likely to be covered by scattered light. Figure 10B The probability graph in the image shows the probability values ​​of the segmented pixels determined in this way.

[0106] Pixels 21 that represent interference with a certain probability can be identified and segmented (and / or binarized) based on probability limit values. For example, all pixels 21 with a probability value equal to or greater than the probability limit value of 0.5 can be identified, and as... Figure 10C The white pixels shown can be set to a value of 1 during segmentation. Remaining pixels in the segmentation with a probability value less than 0.5, and therefore potentially not considered interference, are assigned a value of zero. Figure 10CIn the image, these valid pixels (with a certain probability) are shown in black. Figure 10C The representation in the image can also be described as an image interference mask.

[0107] Alternatively, you can use Figure 9A The distance values ​​in the data can be improved through rationalization. Figure 10C Image interference mask. Since image interference typically has the same (radial) distance value as the object 40 itself, all pixels whose associated distance values ​​lie within a tolerance range (or distance corridor) around the object 40 can be preserved. Figure 10C In the image interference mask, this tolerance range is based on Figure 9A The distance data in the data is determined.

[0108] Then image interference masks can be used. Figure 9A Filter the distance values ​​in the distance data, such as Figure 10D As shown. Then, the distance value of the identified pixel 21 representing interference can be invalidated, and can be set to zero, for example, as... Figure 10D As shown.

[0109] Figure 11A This illustrates information obtained by means of a photoelectric sensor according to an embodiment of the present invention. Figures 9A to 10D A (3D) bird's-eye view representation of distance data for the monitored area (where there are no objects 40). Figure 11B and Figure 11C Each showed Figure 10D The (3D) center perspective or bird's-eye view representation of the filtered distance data, in which pixels 21 identified as interference (as described herein) are highlighted.

[0110] The procedures described herein can be used to identify image interference caused by strongly retroreflective objects, such as retroreflectors. Optoelectronic (industrial) sensors can shield or ignore this (reflector) interference. Furthermore, knowledge of the position, size, and shape of objects (such as retroreflectors) within the field of view of an optoelectronic sensor can be used to induce appropriate system responses, such as automatically adjusting the geometry of the protective field while continuously reducing the travel speed of an autonomous robot. In all cases, the usability of optoelectronic sensors can be improved, which is generally considered an important quality criterion.

Claims

1. A photoelectric sensor for detecting at least one object within a monitored area, in, The photoelectric sensor includes a light emitter, a light receiver, and an evaluation unit; The light emitter is configured to emit transmitted light into the monitoring area. The optical receiver is configured to receive the received light reflected back from the monitoring area, and The evaluation unit is configured as follows: Distance data about the monitoring area is obtained based on the received light, wherein the distance data includes an intensity value and an associated distance value; For all data portions in the distance data that do not originate from the object. In each case, the distance to the object is determined based on the distance data, and, In each case, based on the determined distance between the corresponding data portion and the object, based on the scattered light behavior model of the photoelectric sensor, and based on the intensity value of the corresponding data portion, a probability value is determined that the corresponding data portion represents interference; and All data portions whose determined probability values ​​are equal to or greater than the probability limit are identified as interference.

2. The photoelectric sensor according to claim 1, wherein, The photoelectric sensor is one of a time-of-flight camera and a lidar sensor.

3. The photoelectric sensor according to claim 1, wherein, The evaluation unit is further configured to determine, for all data portions with intensity values ​​less than a first intensity limit, the probability value that the corresponding data portion represents the interference.

4. The photoelectric sensor according to claim 1, wherein, The evaluation unit is configured to determine the distance to the edge of the object based on the distance data in each case.

5. The photoelectric sensor according to claim 1, wherein, The evaluation unit is configured to determine the distance, size, intensity, and / or refraction rate of the reflected object based on the distance data.

6. The photoelectric sensor according to claim 5, wherein, The probability value of the interference represented by the corresponding data portion is determined based on the distance between the corresponding data portion and the object, the intensity value of the corresponding data portion, the maximum measurable intensity, and the size of the object.

7. The photoelectric sensor according to claim 5, wherein, The evaluation unit is configured to determine the size of the retroreflected object based on the portion of the distance data in which the intensity value is equal to or greater than a first intensity limit value.

8. The photoelectric sensor according to claim 5, wherein, The evaluation unit is configured to identify the data portion representing interference in the distance data only if the size of the determined object is equal to or greater than a predetermined size limit, the retroreflectivity of the determined object is equal to or greater than a predetermined second intensity limit, and / or the intensity of the determined object is equal to or greater than a predetermined second intensity limit.

9. The photoelectric sensor according to claim 5, in, The distance data includes multiple pixels, each pixel having an intensity value and an associated distance value; and The evaluation unit is configured as follows: The size of the object is determined based on the pixels in the distance data whose intensity values ​​are equal to or greater than a first intensity limit value. For each pixel whose intensity value is less than the first intensity limit value Determine the distance to the nearest pixel to the object, and Based on the intensity value of the corresponding pixel, the determined distance between the corresponding pixel and the nearest pixel of the object, the maximum measurable intensity and / or the intensity of the object, and the size of the object, the probability value of the corresponding pixel representing interference is determined; and All pixels with a probability value equal to or greater than the probability limit are identified as interference.

10. The photoelectric sensor according to claim 9, in, The size of the object is determined by the number of pixels.

11. The photoelectric sensor according to claim 10, wherein, The distance to the object is determined in pixels.

12. The photoelectric sensor according to claim 1, wherein, The evaluation unit is configured as follows: The distance to the reflected object is determined based on the distance data, and Any data portion whose probability value is equal to or greater than the probability limit value and whose associated distance value is within the tolerance range around the determined object's distance is identified as interference.

13. The photoelectric sensor according to claim 12, in, The distance data includes multiple pixels, each pixel having an intensity value and an associated distance value, and The evaluation unit is configured as follows: Each pixel in the distance data whose intensity value is equal to or greater than the first intensity limit value is input into the distance histogram. Identify the peak value in the distance histogram. The size of the object is determined based on the number of pixels below the peak value, and The distance to the object is determined based on the position of the peak value in the distance histogram.

14. The photoelectric sensor according to claim 13, in, The peak value is the maximum peak value in the distance histogram.

15. The photoelectric sensor according to claim 1, wherein, Data portions identified as interference in the distance data are removed from the distance data, marked as invalid, and / or ignored during further evaluation of the distance data.

16. The photoelectric sensor according to claim 15, wherein, The portion of the distance data identified as interference is used to control the robot's movement.

17. The photoelectric sensor according to claim 1, wherein, The object includes a reflector.

18. The photoelectric sensor according to claim 17, wherein, The reflector is a retroreflector.

19. A system comprising at least one photoelectric sensor and at least one autonomous robot, wherein, The photoelectric sensor includes a light emitter, a light receiver, and an evaluation unit. The light emitter is configured to emit transmitted light into the monitoring area. The optical receiver is configured to receive the received light reflected back from the monitoring area, and The evaluation unit is configured as follows: Distance data about the monitoring area is obtained based on the received light, wherein the distance data includes an intensity value and an associated distance value; For all data portions in the distance data that do not originate from the object. In each case, the distance to the object is determined based on the distance data, and, In each case, based on the determined distance between the corresponding data portion and the object, based on the scattered light behavior model of the photoelectric sensor, and based on the intensity value of the corresponding data portion, a probability value is determined that the corresponding data portion represents interference; and All data portions whose determined probability values ​​are equal to or greater than the probability limit are identified as interference.

20. The system according to claim 19, wherein, The photoelectric sensor is attached to the autonomous robot and can move with it.

21. A method for detecting at least one object in a monitored area, in, Transmitted light is emitted into the monitoring area; Among them, the received light reflected back by the monitored area is received; The distance data about the monitoring area is obtained based on the received light, wherein the distance data includes an intensity value and an associated distance value; Specifically, for all data portions of the distance data that do not originate from the object, In each case, the distance to the object is determined based on the distance data, and, In each case, based on the determined distance between the corresponding data portion and the object, based on the scattered light behavior model of the photoelectric sensor, and based on the intensity value of the corresponding data portion, a probability value is determined that the corresponding data portion represents interference; and Specifically, all data portions in the distance data whose probability values ​​are equal to or greater than a predetermined probability limit are identified as interference.

22. The method according to claim 21, in, The object is either a reflector or a retroreflector.

23. The method according to claim 21, wherein, For all data portions with intensity values ​​less than the first intensity limit, the probability value that the corresponding data portion represents interference is determined.

24. The method according to claim 21, wherein, The determined distance from the object is the distance from the edge of the object.

25. The method according to claim 21, wherein, The distance data for the monitored area was measured using the time-of-flight method.

26. The method according to claim 21, in, The distance to the reflected object is determined based on the distance data; and Specifically, all data portions whose probability values ​​are equal to or greater than the probability limit value and whose associated distance values ​​are within the tolerance range of the determined object's distance are identified as interference.

Images