Method for a detection device; detection device

Abstract

The invention uses a method for a detection device, wherein the detection device is provided to non-destructively acquire a detection signal from an object (16) arranged within an underground examination region (14), as a starting point. It is proposed that the detection signal is assigned to a detection result on the basis of a machine learning process in at least one method step, in order to output the detection result and / or use this to adjust the detection device.

Description

[0001] Description

[0002] Procedure for a

[0003] State of the art

[0004] A method for a detection device has already been proposed, wherein the detection device is intended to non-destructively detect a detection signal from an object arranged within an examination substrate.

[0005] Disclosure of the invention

[0006] The invention is based on a method for a detection device, wherein the detection device is provided to non-destructively detect a detection signal from an object arranged within an examination substrate.

[0007] It is proposed that the detection signal be assigned to a detection result in at least one method step of the method based on a machine learning process in order to output the detection result and / or use it to adjust the detection device. The method preferably comprises an application phase. In the application phase, the detection signal is detected by a sensor unit of the detection device in at least one detection step of the method. In the application phase, the detection signal is evaluated by an evaluation unit in at least one evaluation step of the method. The evaluation unit is preferably a computing unit of the detection device or, alternatively, an external computer system having a data connection, in particular a radio-wave-based connection, with the detection device, in particular an internet server.In the application phase, the detection result is preferably output to a user of the detection device by an output unit of the detection device and / or by an external output device, for example, a smartphone, a tablet, or the like. Additionally or alternatively, in the application phase, the detection result is used by the processing unit of the detection device in a setting step of the method to change a setting of the detection device, in particular to repeat the detection step with a setting of the detection device, in particular of the sensor unit, that is tailored to the object and / or the background.

[0008] The method preferably comprises a learning phase. In the learning phase, sample values ​​(so-called training data) for the detection signal are collected in a data acquisition step of the method. The sample values ​​can be acquired by the sensor unit of the detection device, acquired by a sensor unit of another detection device, generated by a simulation, read from logged data sets, and / or the like. In a learning step of the learning phase, the sample values ​​are processed by the machine learning process into a model that maps the detection signal to the detection result. The machine learning process is preferably carried out by a learning unit. The learning unit is preferably an external computer system, in particular the one already mentioned, alternatively the computer unit of the detection device or another external computer system, for example a private server.Optionally, the method comprises a provision step in which, after the learning phase, the model generated by the machine learning process is transferred to the detection device and stored in a memory of the evaluation unit, in particular the computing unit of the detection device. Preferably, in the application phase, the evaluation unit uses the model generated by the machine learning process to execute the evaluation step. The learning phase can be completed before the application phase or overlap with the application phase, wherein the example values ​​underlying the model are supplemented in particular by the measured values ​​of the detection signal acquired in the application phase. The machine learning process is an algorithm from the field of machine learning, preferably for creating the model based on the example values. The machine learning process preferably comprises a neural network and / or a classification method.The machine learning process can comprise supervised learning and / or unsupervised learning. The method preferably comprises an assignment step in which at least one evaluation parameter is assigned to example values ​​of the detection signal. The detection result can be identical to an evaluation parameter, comprise multiple evaluation parameters, or be determined depending on the evaluation parameters. In supervised learning, the evaluation parameter, in particular the totality of evaluation parameters, for an example value indicates during the learning process to which detection result the model should map this example value. In unsupervised learning, the evaluation parameter is subsequently assigned to a structure in the example values ​​determined by the learning process. The assignment step can be performed manually or automatically during the data acquisition step.The evaluation parameter can, for example, describe the nature of the test substrate, a type of object, a size of the object, a depth of the object in the test substrate, or the like. The evaluation parameter can be known, for example, if an example value for the detection signal is recorded on a known test substrate with a known object. The test substrate and / or the object can be determined, for example, by recreating a detection situation with the detection device in a laboratory or by a computer-based simulation of the detection situation and / or can be taken from a construction plan.The evaluation parameter can, for example, be detected by an external sensor device which comprises other sensors, in particular invasive sensors and / or sensors which cannot be integrated into the detection device due to a limited installation space of the detection device, and / or sensors which are more precise than the sensor unit of the detection device.

[0009] During the application phase, the detection device is particularly intended to be placed against and / or moved against the examination substrate in order to detect the object. "Non-destructive" is understood to mean, in particular, without permanent change, in particular without damage, to the examination substrate and the object. In particular, the detection device detects the object non-invasively. Preferably, in the detection step, the sensor unit generates a signal and transmits it into the examination substrate, in particular as an electromagnetic wave or as an electromagnetic field, wherein the sensor unit detects a change, in particular backscattering, of the signal by the examination substrate and / or the object as a detection signal. Alternatively or additionally, the detection device is intended to detect a signal emanating from the object as a detection signal.The detection device is intended, for example, to detect a metallic object, a non-metallic object, an electrical line, in particular a low-voltage line, a single-phase alternating current line or a three-phase current line, a wooden beam, a metal support, or a plastic pipe, in particular a water-filled plastic pipe or a non-water-filled plastic pipe, or the like. The surface to be examined is, for example, a wall, a ceiling, a floor, or a piece of furniture in a building.

[0010] The inventive design advantageously allows the detection signal to be evaluated in detail. In particular, additional information contained in the detection signal, which goes beyond the mere presence or absence of the object, can advantageously be reliably extracted from the detection signal. By utilizing the additional information, operation of the detection device can be advantageously kept simple and / or a user of the detection device can be provided with an advantageously large amount of information, in particular with only a single measurement with the detection device. Furthermore, an advantageously reliable repeatability of the detection result can be achieved and, in particular, a dependence of the detection result on a travel path of the detection device on the examination surface can advantageously be kept low.

[0011] It is further proposed that in a learning phase of the machine learning process, in particular the one already mentioned, example values ​​for the detection signal are divided into different groups depending on a clustering of the example values. The learning phase preferably comprises a preprocessing step in which the learning unit prepares the example data before the learning step. In the preprocessing step, the learning unit preferably carries out unsupervised learning in order to divide the example values ​​into the groups. In the preprocessing step, the learning unit preferably divides the example values ​​into at least two different groups. The learning unit determines the groups and the affiliation of the example values ​​to these groups, preferably by means of a clustering method. In the preprocessing step, the learning unit preferably extracts at least one feature, preferably a plurality of features, from each example value to be processed.The feature can, for example, be a physical quantity, a statistical quantity, or an abstract key figure of the example value. If example values ​​have similar values ​​for the feature / all features, the learning unit preferably assigns the example values ​​to the same group. If example values ​​have different values ​​for the feature / one of the features, the learning unit preferably assigns the example values ​​to different groups. Whether the values ​​of the features are similar or different is defined by the clustering method used. The embodiment according to the invention allows example values ​​that are similar to one another based on the extracted features to be summarized and processed jointly. In particular, the same evaluation parameters can be assigned to all example values ​​within the same group with an advantageously low risk of error.In particular, the time required to execute the assignment step can advantageously be kept short. In particular, an example value acquired in an unknown acquisition situation can be assigned an evaluation parameter of an example value of the same group acquired in a known acquisition situation.

[0012] It is further proposed that the machine learning process, particularly in the preprocessing step, uses nearest-neighbor classification to perform a division into different groups. Particularly preferably, the learning unit uses a K-Nearest-Neighbor algorithm (KNN) to perform the division into the different groups. In particular, the learning unit assigns an example value to a group depending on which group the majority of neighboring example values ​​have been assigned to. Neighboring example values ​​are considered to be a specified number of example values ​​that, in a parameter space spanned by the features, have the smallest distance to the example value to be classified. The distance between two example values ​​in the parameter space spanned by the features can be determined using the Euclidean metric, the Manhattan metric, or another metric.The number of neighboring sample values ​​used can be specified by an operator of the learning unit and / or determined using an optimization algorithm. Optionally, in the case of a majority decision among the neighboring sample values, a share of the votes of the sample values ​​is weighted, in particular with the respective distance of the neighboring sample values ​​from the sample value to be classified. Alternatively, the sample values ​​are classified into different groups using a cuboid classifier, a distance classifier, a polynomial classifier, or another classifier. The inventive design enables an advantageously reliable classification of the sample values ​​into the groups.

[0013] It is further proposed that, in a learning phase of the machine learning process, an evaluation parameter, in particular the evaluation parameter(s) already mentioned, be assigned to the example values ​​for the detection signal in groups. Preferably, at least one example value is selected from each group as a representative for the group. The representative is an example value whose evaluation parameters are known through replication, simulation, or from other sources. Preferably, the learning unit assigns the evaluation parameter(s) assigned to the representative to all example values ​​within the representative's group. The representative can be added to the remaining example values ​​in the data acquisition step or determined after the remaining example values ​​have been divided into groups.Optionally, the learning unit uses the representatives and / or domain knowledge as support points for the nearest neighbor classification. Alternatively, the number of groups is determined automatically, for example by varying the initial cluster centers and minimum sum distances. The embodiment according to the invention advantageously allows the duration of the assignment step to be kept short. In particular, an advantageously large number of example values ​​can be used for the machine learning process, the evaluation parameters of which have not been recorded individually. Furthermore, it is proposed that in a learning phase of the machine learning process, in particular the one already mentioned, a data set of example values ​​for the detection signal consists at least substantially of user data. “Substantially” should be understood to mean in particular more than 30%, preferably more than 60%, particularly preferably more than 90%, extremely preferably more than 99%.Preferably, the data set comprises the representatives of the groups in addition to the user data. "User data" should be understood in particular as detection signals that are detected by a user in the application phase with the detection device or another detection device, in particular in a real, i.e., not recreated or simulated, detection situation. User data is preferably collected from a plurality of detection devices to form the data set. In particular, the user data is automatically transmitted from a data interface of the detection device to the learning unit. The embodiment according to the invention advantageously allows the time and cost required for recreated and / or simulated detection situations to be kept low.Furthermore, the sample values ​​advantageously include many real-world, not just ideal, acquisition situations, which, particularly due to their complexity and / or diversity, cannot be replicated in the laboratory. In particular, a robust model can be created using the machine learning process.

[0014] Furthermore, it is proposed that, in at least one method step, an operating mode of the detection device is automatically selected depending on the machine learning process. The detection device, in particular the sensor unit, comprises at least one operating mode provided for a specific detection situation and at least one further operating mode provided for a further detection situation. Detection situations differ, for example, due to different materials or constructions of the examination substrate and / or due to different types of detected object. The operating modes differ, for example, due to a sensor element used by the sensor unit and / or due to a signal parameter of the emitted signal, for example intensity, frequency, duration, or the like.Preferably, the sensor unit detects an initial detection signal in a detection situation, which is evaluated by the evaluation unit to produce an initial result. Preferably, the evaluation unit, in particular the computing unit of the detection device, selects one of the operating modes based on the initial result in order to carry out all subsequent detection steps in the same detection situation in the selected operating mode. Preferably, a table is stored in a memory of the evaluation unit, in particular the computing unit of the detection device, which table assigns the operating modes to the initial result and is read out by the evaluation unit to select the operating mode. Alternatively, the detection result includes the operating mode intended for the detection situation as an evaluation parameter and is learned during the machine learning process.The inventive design allows for advantageously simple and intuitive operation of the detection device. In particular, user adjustment of the detection device is eliminated. In particular, the number of required operating elements of the detection device and / or menu items of a graphical interface of the detection device can be advantageously kept low. In particular, the risk of incorrect operation of the detection device by a user can be advantageously kept low.

[0015] It is further proposed that, in at least one method step, a zero reference for the detection signal be automatically selected as a function of the machine learning process. The zero reference is a measured value of the detection signal when the examination substrate is free of detectable objects. Preferably, the evaluation unit evaluates a difference between the detection signal and the zero reference to determine whether or not an object is present in the examination substrate. The zero reference normally depends on the material and construction of the examination substrate. In particular, the different operating modes of the detection device have different zero references. The zero reference(s) is / are preferably part of the model created by the machine learning process. In particular, in the assignment step, a zero reference is assigned to each example value, in particular to each group of example values.Alternatively, the zero references are stored as a table in a memory of the processing unit and are read by the processing unit depending on the selected operating mode. The inventive design eliminates the need for a user of the detection device to separately determine the zero reference. In particular, the detection device can advantageously be used quickly to detect an object after being switched on. In particular, the risk of incorrect operation of the detection device can be advantageously minimized.

[0016] It is further proposed that the detection device, in at least one method step, outputs an evaluation parameter, in particular the one already mentioned, of a group of example values ​​of the detection signal, in particular the one already mentioned, when a measured value of the detection signal is assigned to this group. The evaluation unit preferably assigns the evaluation parameter to the measured value using the model. The output unit preferably outputs at least whether an object is detected or whether no object is detected. The output unit preferably outputs the evaluation parameter in addition to the information as to whether an object is detected or not. For example, the output unit outputs a material or a construction method of the examination substrate as an evaluation parameter.For example, the output unit outputs the type of detected object, the size of the detected object, in particular a maximum extension of the detected object parallel to a surface of the examination substrate, a distance of the detected object from the surface of the examination substrate, a distance, in particular parallel to the surface of the examination substrate, between an edge of the detected object and a current measurement point of the detection device, and / or the like, as evaluation parameters. The inventive design advantageously allows a user to be provided with a large amount of information, particularly with a single measurement.

[0017] Furthermore, a detection device is proposed with at least one sensor unit, in particular the one already mentioned, for detecting a detection signal, in particular the one already mentioned, and with at least one computing unit, in particular the one already mentioned, for carrying out a method according to the invention. The sensor unit comprises at least one sensor element. Optionally, the sensor unit comprises a plurality of, in particular differently designed, sensor elements. Examples of sensor elements of the sensor unit include a radar, in particular a narrowband radar, in particular in a frequency range from 2.4 GHz to 2.4835 GHz, and / or an ultrawideband radar, an inductive sensor, a capacitive sensor, an alternating current sensor, in particular a 50 Hz alternating current sensor and / or a 60 Hz alternating current sensor, or the like.A "computing unit" is understood, in particular, to be a unit with an information input, an information processing unit, and an information output. The computing unit advantageously comprises at least one processor, a memory, input and output means, further electrical components, an operating program, control routines, control routines, and / or calculation routines. The components of the computing unit are preferably arranged on a common circuit board and / or advantageously arranged in a common housing. The detection device preferably comprises the output unit for outputting the detection result. The output unit comprises, for example, a display, a microphone, a vibration alarm, or the like for outputting the detection result. The detection device preferably comprises the data interface.The data interface can comprise a wired and / or radio-wave interface element, in particular a Bluetooth interface, a WLAN interface, an Ethernet interface, or the like. The data interface is provided for outputting the detection result to the external output device, for forwarding the detection signal to the external computer system, and / or for receiving the model from the external computer system. The inventive design makes it possible to provide a detection device that is advantageously easy to operate and / or that advantageously provides a large amount of information.

[0018] The method according to the invention and / or the detection device according to the invention should not be limited to the application and embodiment described above. In particular, the detection device according to the invention and / or the method according to the invention may, in order to fulfill a function described herein, have a number of individual elements, components, units, and method steps that differs from the number stated herein. Furthermore, in the value ranges specified in this disclosure, values ​​within the stated limits are also to be considered disclosed and can be used arbitrarily. Drawings

[0019] Further advantages will become apparent from the following description of the drawings. The drawings illustrate an exemplary embodiment of the invention. The drawings, the description, and the claims contain numerous features in combination. Those skilled in the art will also expediently consider the features individually and combine them into useful further combinations.

[0020] They show:

[0021] Fig. 1 is a schematic representation of a detection device according to the invention,

[0022] Fig. 2 is a schematic flow diagram of a method according to the invention and

[0023] Fig. 3 shows a schematic division of a data set of example data in the course of a machine learning process of the method according to the invention.

[0024] Description of the embodiment

[0025] Figure 1 shows a detection device 12. The detection device 12 is shown arranged on a surface of an examination substrate 14, for example, a wall. The detection device 12 is intended to non-destructively detect a detection signal from an object 16 arranged within the examination substrate 14. The detection device 12 is preferably designed to be hand-held, preferably held with one hand, and in particular operable with the same hand. Particularly preferably, the detection device has a total volume of less than 7000 cm 3 , preferably less than 5000 cm 3 , particularly preferably less than 3000 cm 3The detection device 12 comprises a sensor unit 24 for detecting the detection signal. The sensor unit 24 preferably comprises at least one sensor element designed to emit an electromagnetic signal into the examination substrate 14 and to detect a portion of the emitted signal scattered back by the object 16 as a detection signal. Alternatively or additionally, the sensor unit 24 comprises a capacitive sensor and / or an inductive sensor for detecting the detection signal. The detection device 12 comprises a computing unit 26 for evaluating the detection signal to produce a detection result.

[0026] The detection device 12 preferably comprises an output unit 28 for outputting the detection result. The output unit 28 preferably comprises a display for displaying the detection result. The detection device 12 preferably comprises at least one data interface 32 for exchanging data with an external computer system and / or an external output device. The detection device 12 preferably comprises a power supply unit 34 for supplying the sensor unit 24, the computing unit 26, the output unit 28, and / or the data interface 32 with electrical energy. The power supply unit 34 is provided, in particular, for accommodating at least one battery and / or an accumulator.The detection device 12 preferably comprises at least one operating element 30, for example a switch, a button, a slider, or the like, for operating the detection device 12, in particular for triggering a detection with the detection device 12. The detection device 12 comprises a housing 36, within which at least the sensor unit 24 is arranged. The computing unit 26, the data interface 32, and / or the power supply unit 34 are preferably arranged in the housing 36. The output unit 28 and / or the operating element 30 are preferably arranged on the housing 36 or embedded in the housing 36. The external computing system and / or the external output device are / is preferably designed independently of the detection device 12 (not shown here).

[0027] Figure 2 shows a flowchart of a method 10 for the detection device 12. The method 10 preferably comprises a detection step 40, in which the sensor unit 24 detects the detection signal. The method 10 comprises an evaluation step 42, in which the detection signal is evaluated as a function of a machine learning process to obtain the detection result. The method 10 particularly comprises an output step 44, in which the detection result is output. The method 10 preferably comprises a setting step 46, in which the computing unit 26 changes a setting of the detection device 12, in particular of the sensor unit 24 and / or the computing unit 26, as a function of the detection result.

[0028] The method 10 preferably comprises an application phase 38. The application phase 38 is preferably carried out completely, alternatively partially, by means of the detection device 12. The application phase 38 preferably comprises the detection step 40, the evaluation step 42, the output step 44 and / or the setting step 46. In particular, the evaluation step 42 is carried out by the computing unit 26. Alternatively, the evaluation step 42 is carried out by an external computing system, in particular the one already mentioned or a further external computing system. The output step 44 is preferably carried out by the output unit 28. Alternatively, the output step 44 is carried out by the external output device. In particular, in the alternative embodiments, the data interface 32 transmits the detection signal to the further external computing system, receives the detection result from the external computing system and / or sends the detection result to the external output device.

[0029] The method 10 preferably comprises a learning phase 18. The learning phase 18 is preferably performed by the external computer system, alternatively by the computing unit 26. In the learning phase 18, a model is created using the machine learning process, which maps the detection signal to the detection result. The model is stored in a provision step 58 in a memory of the computing unit 26 of the detection device 12. In particular, the computing unit 26 applies the model in the evaluation step 42 to convert a measured value of the detection signal acquired by the sensor unit 24 into the detection result.

[0030] The learning phase 18 preferably comprises a data acquisition step 48. In the data acquisition step 48, sample values ​​22 (cf. Fig. 3) are acquired, which are processed to create the model. The method 10 preferably comprises a collection step 56, in which measured values ​​of the detection signal acquired by the sensor unit 24 are collected as sample values ​​22. The data interface 32 preferably transmits the measured values ​​acquired by the sensor unit 24, in particular collectively or individually, to the external computer system in the collection step 56. Particularly preferably, the external computer system collects measured values ​​from a plurality of detection devices, in particular within the framework of an Internet of Things (IoT) concept. A data set comprising the sample values ​​22 for the detection signal consists at least substantially of user data.User data are measured values ​​of the detection signal that are acquired by users in the application phase 38 using the detection device 12. Preferably, in the data acquisition step 48, the manufacturer of the detection device 12 additionally determines example values ​​22 as representatives. The representatives are determined, for example, by acquiring data with the detection device 12 on a replica of various examination surfaces 14, including different objects 16, and / or by simulating acquisition steps 40 on various examination surfaces 14, including different objects 16. Preferably, the data set comprises at least 10 times, preferably at least 100 times, particularly preferably at least 1000 times, more user data than representatives.

[0031] The learning phase 18 preferably comprises a preprocessing step 50. In the preprocessing step 50, the example values ​​22 are evaluated with respect to at least one feature 60, 62, preferably a plurality of features 60, 62 (see Fig. 3). The at least one feature 60, 62 is preferably provided as a distinguishing criterion, based on which the external computer system evaluates whether two example values ​​22 are the same or different. Features 60, 62 of the example values ​​can be, for example, an intensity of the detection signal, a ratio of the backscattered portion to the emitted portion of the detection signal, a mean value of the detection signal, an extreme value of the detection signal, a duration of an amplitude modulation of the detection signal caused by the object, a fluctuation range of the detection signal, or the like.The example values ​​22 for the detection signal are divided into different groups 20 by the external computer system depending on a cluster formation of the example values ​​22. The external computer system evaluates the cluster formation of the example values ​​22 preferably based on the determined features 60, 62. In particular, the external computer system determines whether a distribution of the example values ​​22 within a parameter space spanned by the at least one feature 60, 62 has clusters. A dimensionality of the parameter space is equal to the number of different features 60, 62 per example value 22. A distribution of the example values ​​22 in a two-dimensional parameter space spanned by a feature 60 and a further feature 62 is shown by way of example in Figure 3. The external computer system preferably assigns the example values ​​22 forming a cluster to the same group 20.Preferably, the external computing system assigns sample values ​​22 that form different clusters to different groups 20. In Figure 3, the sample values ​​22 were divided into ten groups 20 as an example. The external computing system uses a nearest-neighbor classification to divide the sample values ​​22 into the different groups 20.

[0032] The method 10 preferably comprises an assignment step 52. In the assignment step 52, at least one evaluation parameter is assigned to the example values ​​22 for the detection signal in groups. Preferably, at least one evaluation parameter is assigned to each group 20. Optionally, several evaluation parameters are assigned to at least one group 20. Two different groups 20 differ from one another by at least one of the evaluation parameters assigned to them. The at least one evaluation parameter is processed by the machine learning process, in particular as an intermediate result to be learned or as a final result to be learned. Preferably, a background parameter that describes or characterizes the examination background 14 is assigned to the groups 20 as an evaluation parameter.The subsurface parameter includes, for example, the following values: concrete, lightweight construction, drywall, in particular the thickness and / or number of plasterboard panels and / or wooden panels of lightweight construction or drywall, brick wall, in particular the type of stone of the brick wall, the presence of underfloor heating or wall heating, or the like. Preferably, an object type parameter that describes or characterizes the type of object 16 is assigned to the groups 20 as an evaluation parameter. The object type parameter includes, for example, the following values: metal, non-metal, current-carrying, non-current-carrying, magnetic, non-magnetic, low-voltage cable, single-phase alternating current cable, in particular between 110 V and 230 V, multi-phase alternating current cable, three-phase cable, wooden beam, metal support, plastic pipe, water-filled pipe, in particular freshwater pipe, non-water-filled pipe, in particular sewage pipe.Preferably, a depth parameter is assigned to the groups 20 as an evaluation parameter, which describes or characterizes a distance of the object 16 from a surface of the examination substrate 14. The depth parameter can specify the distance continuously or in regions. Preferably, a shape parameter is assigned to the groups 20 as an evaluation parameter, which describes or characterizes a shape or extent of the object 16. The shape parameter specifies, for example, a diameter of the object 16. The assignment step 52 is performed by the external computer system using the representatives.The evaluation parameters of the representatives can be recorded by the manufacturer of the detection device 12, for example, using a sensor device on the replicas that is independent of the detection device 12, read from data sheets for the replicas, or read from a simulation program and stored in a memory of the external computer system. Preferably, the external computer system determines at least one representative for each group 20, reads its evaluation parameters from the memory of the external computer system, and assigns these evaluation parameters to all example values ​​22 in the same group 20 as the representative. If there are multiple representatives with different evaluation parameters in the same group 20, the external computer system optionally further subdivides these groups 20 into subgroups.

[0033] The learning phase 18 comprises, in particular, a learning step 54. In the learning step 54, the external computer system creates the model using the machine learning process. In particular, the external computer system uses the example values ​​22 as the model's input values ​​and the evaluation parameters assigned to the example values ​​22 as the model's output value. The detection result is composed, in particular, of the entirety of the evaluation parameters or is determined depending on the evaluation parameters. In the provision step 58, the external computer system transmits the model via the data interface 32 to the computing unit 26 of the detection device 12.

[0034] In the application phase 38, the computing unit 26 of the detection device 12 uses the model to automatically select an operating mode of the detection device 12. Preferably, the sensor unit 24 detects an initial detection signal in the detection step 40, which is evaluated in the evaluation step 42 to produce an initial result. The computing unit 26 uses the model to determine, in particular, the background parameter as a function of the initial detection signal in order to automatically select a zero reference for the detection signal in the setting step 46 as a function of the machine learning process. The computing unit 26 uses the model to determine, in particular, the object type parameter, the depth parameter, and / or the shape parameter as a function of the initial detection signal in order to increase the sensitivity of the sensor unit 24 for the object 16 in the setting step 46 and / or to avoid oversaturation of the sensor unit 24.

[0035] In the application phase 38, in the output step 44, the detection device 12 outputs the at least one evaluation parameter of the group 20 of example values ​​22 of the detection signal if the model assigns a measured value of the detection signal to this group 20. The output unit 28 preferably outputs whether the object 16 was detected. The output unit 28 preferably outputs the object type parameter and the background parameter. The output unit 28 optionally outputs the depth parameter and / or the shape parameter.

Claims

Claims 1. Method for a detection device, wherein the detection device is designed to non-destructively detect a detection signal from an object (16) arranged within an investigation substrate (14), characterized in that the detection signal is assigned to a detection result in at least one method step on the basis of a machine learning process in order to output the detection result and / or to use it for setting the detection device.

2. Method according to claim 1, characterized in that in a learning phase (18) of the machine learning process, example values ​​(22) for the detection signal are divided into different groups (20) depending on a clustering of the example values ​​(22).

3. Method according to claim 2, characterized in that the machine learning process uses a nearest-neighbor classification to perform a division into different groups (20).

4. Method according to one of the preceding claims, characterized in that in a learning phase (18) of the machine learning process, at least one evaluation parameter is assigned to example values ​​(22) for the detection signal in groups.

5. Method according to one of the preceding claims, characterized in that in a training phase (18) of the machine learning process a data set of example values ​​(22) for the detection signal consists at least substantially of user data. A method according to any one of the preceding claims, characterized in that, in at least one method step, an operating mode of the detection device is automatically selected depending on the machine learning process. A method according to any one of the preceding claims, characterized in that, in at least one method step, a zero reference for the detection signal is automatically selected depending on the machine learning process. A method according to any one of the preceding claims, characterized in that, in at least one method step, the detection device outputs an evaluation parameter of a group of example values ​​(22) of the detection signal when a measured value of the detection signal is assigned to this group (20).A method according to any of the preceding claims, characterized in that the evaluation parameter is assigned at least one background parameter, at least one object parameter, at least one depth parameter and / or at least one shape parameter. A detection device comprising at least one sensor unit (24) for acquiring a detection signal and at least one processing unit (26) for carrying out a method according to any of the preceding claims.

Images