Method and system for detecting suspicious objects in soil

Abstract

The invention relates to a method and a system for detecting suspected objects, in particular munitions, in soil. The method (100) comprises: creating (110) at least three exploratory boreholes in soil; performing (120) at least one measurement within the exploratory boreholes using a sensor; generating (130) measurement data based on the measurements performed; analyzing (140) the measurement data using an algorithm trained on the basis of a training dataset to categorize an area of ​​soil spanned by the exploratory boreholes; categorizing (150) the spanned area, depending on the measurement data, using the algorithm into one of at least two categories, comprising a first category according to which the measurement data indicate that no suspected object is present and a second category according to which the measurement data indicate that a suspected object is present.Carry out (160) further investigations to determine whether a suspected object is present in the area covered if the area covered by the exploratory drilling has not been categorized in the first category.;

Description

[0001] The invention relates to a method for detecting suspected objects, in particular munitions, in soil. Furthermore, the invention relates to a system for detecting suspected objects, in particular munitions, in soil.

[0002] Traditional ordnance detection methods can be used to search for unexploded ordnance. This ordnance can include unexploded ordnance such as bombs, ammunition, or any type of wartime debris. If such ordnance is located through ordnance detection, it can be recovered and removed as part of ordnance disposal procedures.

[0003] To detect suspected objects, especially unexploded ordnance, in the ground, boreholes, known as exploratory boreholes, are typically created using drilling equipment. After creating such a borehole, a probe tube is usually inserted into it. Measurements can then be taken using measuring instruments that are guided through the probe tube.

[0004] To successfully detect a suspected object, sufficient contrast between the object's characteristics and its environment, as well as a sufficient signal-to-noise ratio, are required. This means that the signal generated by the suspected object must be sufficiently strong to outweigh any interfering signals and other artifacts present at the measurement location in order to detect it.

[0005] One such contrast in properties, for example in physical characteristics, between munitions, such as unexploded bombs, and the surrounding soil lies particularly in magnetizability. Since munitions like unexploded bombs typically contain steel, they usually exhibit a significantly higher magnetizability and intrinsic magnetism than the surrounding soil. Therefore, magnetometers are commonly used for the detection of unexploded bombs.

[0006] One problem with detecting suspected objects like unexploded bombs is that in many soils, for example in residential areas or on former military sites, measurements are affected by relatively strong magnetic background noise, also referred to above as an artifactual signal. Background noise occurs to a greater or lesser degree in every measurement process. Background noise refers specifically to interference that overlays and impairs the useful signal of a measurement. Such background noise can have various causes, such as deposits with ferromagnetic components, for example in the form of settlement layers, rubble, construction debris, industrial slag, household and commercial waste, or external electromagnetic interference sources such as radio transmitters, overhead power lines, large moving magnetic objects, and switching of power in electrical installations.

[0007] German patent application DE 20 2022 105 220 U1 discloses a system for detecting landmines in a specific geographical area. The system comprises: a vehicle configured to move within that geographical area; and a landmine detection unit configured to be mounted on the vehicle. The landmine detection unit includes: an ultrasonic sensor used to detect obstacles in the vehicle's path and to identify landmines; a metal detection unit comprising a metal detection coil to detect metal in the vehicle's path; and a thermal imaging camera located on top of the vehicle to obtain object details, including the temperature, of obstacles and metals in the vehicle's path.

[0008] From DE 10 2021 122 140 A1, a method for determining the location of unexploded ordnance in the ground is known, wherein a test borehole is drilled into the ground, and a probe designed to detect anomalies in the ground is inserted into the test borehole, and the positions of the individual test boreholes are determined and stored using a GPS antenna. A drilling rig is provided, which includes a drill bit designed for creating the test borehole, and a GPS antenna is positioned relative to the drill bit and fixed in that position, and during the creation of the test borehole, the position of the drill bit, and thus of the test borehole, is determined using the GPS antenna.

[0009] From DE 10 2018 128 962 A1 a method for determining the position of an object is known, comprising at least one nonlinear component, in particular a semiconductor component, which, when irradiated with high-frequency transmission signals from at least two different positions, generates and re-emits object signals with twice and / or three times the frequency of the respective transmission signal.

[0010] From EP 4 130 404 A1, an excavator for removing soil in which a problematic object, such as unexploded ordnance, may be present is known. The excavator includes a bucket which incorporates a measuring device with magnetometers arranged at intervals between them, configured to measure the local magnetic field vector.

[0011] The spacing between adjacent exploratory boreholes can be larger the lower the background noise and the larger the suspected objects being sought and the resulting useful signal. Basic formulas for determining the appropriate borehole spacing were described, for example, by Wegener and Fleischmann in the article "Location of Deep-lying Unexploded Bombs," Zeitschrift für angewandte Physik, 6.3 (1954), pp. 120-127. This article describes how noise signals can be determined for measurements in exploratory boreholes.These basic calculation formulas for determining borehole spacing also form the basis of current regulations, for example those of the Hanseatic City of Hamburg, as described in the TA-KRD Hamburg 2017 (Technical Instruction for the execution of tasks of systematic searching and uncovering suspected objects / munitions, Free and Hanseatic City of Hamburg, Fire Department, Explosive Ordnance Disposal Service Hamburg).

[0012] The spacing between boreholes, and thus the borehole grid, is generally designed to ensure an overlap of the circles defined by detection radii between two adjacent boreholes, depending on the local signal-to-noise ratio. This is typically achieved if the distance between any two adjacent boreholes does not exceed 1.5 m. However, particularly in areas with relatively high background noise, it may be necessary to reduce the spacing between boreholes, at least locally, to guarantee an overlap of the circles defined by detection radii.

[0013] The selected borehole spacing between adjacent exploratory boreholes must therefore be checked and adjusted if necessary to ensure overlap—and thus reliable and complete detection of suspected objects. However, such a check and adjustment of the borehole spacing is relatively complex. Firstly, the evaluation itself is typically quite time-consuming. Secondly, the process of analyzing whether the selected spacing between the boreholes was sufficient to completely cover the entire area after drilling and evaluating the measurement data is also relatively time-consuming. In certain locations where coverage—i.e., an overlap of the circles defined by the detection radii—is not achieved, additional exploratory boreholes are subsequently drilled to ensure adequate coverage in these areas as well.

[0014] The invention is therefore based on the objective of providing an improved solution that addresses the aforementioned problems. In particular, the object of the invention is to provide a solution by means of which reliable detection of suspected objects, especially munitions, in soil is enabled with less effort, and especially less time.

[0015] According to a first aspect, this problem is solved by a method for detecting suspected objects, in particular munitions, in soil with the features of claim 1. The method comprises: creating at least three, in particular exactly three, spaced-apart probe boreholes in soil, wherein the at least three probe boreholes form a borehole arrangement, preferably a borehole triplet, wherein preferably the at least three probe boreholes of the borehole arrangement each have a probe borehole axis that is arranged parallel to each other and / or vertically aligned; performing at least one physical measurement within each of the at least three probe boreholes of the borehole arrangement by means of at least one sensor;Generating measurement data based on at least one physical measurement performed for each of the at least three exploratory boreholes of the borehole array; analyzing the measurement data associated with the borehole array using an algorithm trained on a training dataset to categorize an area of ​​the ground spanned by the exploratory boreholes of the borehole array; categorizing the area of ​​the ground spanned by the exploratory boreholes of the borehole array, depending on the measurement data associated with the borehole array, using the algorithm into one of at least two categories, wherein the at least two categories comprise: a first category according to which the measurement data indicate that no suspected object is present in the area of ​​the ground spanned by the exploratory boreholes of the borehole array;a second category, according to which the measurement data indicate that a suspected object is present in the area of ​​the ground spanned by the exploratory boreholes of the borehole arrangement; carrying out further investigations to determine whether a suspected object is present in the area of ​​the borehole arrangement, in particular in the area of ​​the ground spanned by the exploratory boreholes of the borehole arrangement, if the area of ​​the ground spanned by the exploratory boreholes of the borehole arrangement has not been categorized in the first category, in particular if the area of ​​the ground spanned by the exploratory boreholes of the borehole arrangement has been categorized in the second category.

[0016] A method is proposed in which several, preferably three, exploratory boreholes, forming a borehole array, are created in the ground. A physical measurement is performed in each of these boreholes using at least one sensor. The measurement data obtained are analyzed using an algorithm, and the area of ​​the ground spanned by the boreholes of the array is categorized by the algorithm, which has been previously trained using a training dataset. At least two categories are provided for this purpose. According to the first category, the corresponding measurement data indicate that no suspected object is present in the area of ​​the ground spanned by the boreholes of the array.According to the second category, the corresponding measurement data indicate that a suspected object is present in the area of ​​the ground exposed by the exploratory boreholes of the borehole array. If the area of ​​the ground exposed by the exploratory boreholes of the borehole array was not categorized in the first category, for example, if the area of ​​the ground exposed by the exploratory boreholes of the borehole array was categorized in the second category, further investigations are carried out to determine whether a suspected object is present in the area of ​​the borehole array, in particular in the area of ​​the ground exposed by the exploratory boreholes of the borehole array.

[0017] One advantage of this method is that, based on the analysis and categorization of the measurement data using the algorithm, it becomes apparent relatively quickly after the data is generated whether or not a suspected object might be present in a specific area of ​​the soil. This very fast and automated categorization of soil areas allows for a much faster and more comprehensive search for suspected objects in the area under investigation. It also makes it possible to check immediately after categorizing a soil area whether the area in question has been assigned to the first category, and in such a case, to conduct further investigations immediately.

[0018] Another advantage is that, due to the immediate and automated categorization, a technician performing exploratory drilling using a dedicated device can be immediately alerted to potentially hazardous areas – specifically, those identified by the categorization. This enables the technician to carry out further processing steps in and around such potentially hazardous areas with particular care in order to minimize hazards, especially the risk of explosion.

[0019] The soil in which the process is carried out can, in particular, be soil on land or soil at the bottom of a body of water.

[0020] It is preferred that the process steps of the procedure are carried out in the specified order. Preferably, the process steps of the procedure are carried out several times in succession for several, in particular adjacent, borehole arrangements, preferably until a soil to be investigated has been completely searched for suspected objects.

[0021] The borehole arrangement consists of at least three, preferably exactly three, spaced-apart probe holes. If exactly three probe holes are provided, they form a borehole arrangement in the form of a triple borehole. The probe hole axes preferably extend vertically, but it is also possible for the probe hole axes to be oriented slightly differently from the vertical. A borehole arrangement preferably comprises three probe holes. A probe hole can also be associated with several borehole arrangements, in particular adjacent probe holes. However, two borehole arrangements must differ in at least one probe hole to be considered different borehole arrangements.

[0022] The generation and analysis of measurement data is preferably carried out by means of at least one data processing unit, preferably several data processing units. The at least one data processing unit is preferably connected to the at least one sensor via a signal connection, so that the sensor data can be transmitted from the at least one sensor to the at least one data processing unit.

[0023] The measurement data associated with the borehole array are analyzed using an algorithm trained on a training dataset. An algorithm trained on a training dataset refers specifically to an artificial intelligence algorithm that has been trained using such a dataset. The analysis of the measurement data and categorization of the soil area exposed by the boreholes of the array is therefore performed using such an AI algorithm.

[0024] The analysis and categorization can be performed as follows: The data within a borehole array, in the form of a borehole triplet, is pre-categorized using artificial intelligence. A so-called random forest algorithm is trained with approximately 80,000 borehole array data points, corresponding to approximately 200,000 borehole triplets previously generated in the soil. Various key figures are determined from the individual components of the measurement data and used as a data basis together with the geometric values ​​of the individual exploratory boreholes and borehole arrays. The initial categorization of the borehole arrays was previously carried out by a large number of experts to ensure the objectivity of the categorization. For the categorization, or rather...For example, three categories can be defined for classification, which are explained below as the first, second, and third categories. This categorization or classification serves, in particular, the purpose of identifying areas and / or regions of the ground where the presence of a suspected object, such as an unexploded ordnance, cannot be ruled out. Further work steps are then necessary in these areas. The terms categorization and classification are used synonymously in this document.

[0025] As an alternative or supplement to using a random forest algorithm, an autoencoder can be employed to automatically search for parameters. These parameters do not necessarily have to have a direct physical relationship, but they are very effectively utilized by the algorithm. The parameters determined by the autoencoder can then be used, for example, as a data basis in a so-called XG-Boost algorithm.

[0026] The categorization of the soil area spanned by the exploratory boreholes of the borehole array is carried out using the algorithm, based on the measurement data assigned to the borehole array. The at least two categories comprise: a first category and a second category, and optionally a third category. Preferably, in the case of categorization into the first category, the indicators obtained from the measurement data suggest that no suspected object, in particular no unexploded bomb, is present within the borehole array and / or that a suspected object, in particular an unexploded bomb, can be ruled out within the borehole array.Preferably, in the case of categorization into the second category, the key figures obtained from the measurement data indicate that a suspected object, in particular an unexploded bomb, could be present within the drilling arrangement and / or that a suspected object, in particular an unexploded bomb, cannot be ruled out within the drilling arrangement.

[0027] Further investigations are carried out in the area of ​​the borehole array, in particular, if the area of ​​the ground exposed by the exploratory boreholes of the borehole array has not been categorized in the first category, i.e., if it cannot be ruled out by automated categorization that a suspected object, especially an unexploded bomb, is present within the borehole array. These further investigations may include, in particular, creating additional exploratory boreholes in the area of ​​the borehole array, especially within the area exposed by the exploratory boreholes assigned to the borehole array.

[0028] Preferably, categorization is also possible at different depths, particularly in the vertical direction. This allows for the determination of whether further investigations should or can focus on one or more specific depth ranges. Such depth-dependent analysis and categorization makes it particularly advantageous to ensure that further investigations are carried out in a targeted manner at specific depths or depth ranges. This, in particular, reduces the time required for further investigations.

[0029] According to a particularly preferred embodiment, the categorization by means of the algorithm is carried out into one of at least three categories, wherein the at least three categories comprise: the first category, the second category, and a third category, according to which, in particular due to interference signals, a categorization into the first category or into the second category is not reliably possible.

[0030] Preferably, when categorized into the third category, the key figures obtained from the measurement data do not directly point to a suspected object, in particular an unexploded bomb, within the borehole arrangement. However, categorization into the first or second category is not reliably possible because the measurement data are, for example, so strongly superimposed by magnetic interference signals that the key figures obtained from the measurement data are so distorted or are not clearly distinguishable from the interference signals or do not clearly exceed the interference signal level that a potential suspected object, in particular an unexploded bomb, in the borehole arrangement would not be reliably detected.

[0031] Preferably, further investigations are carried out to determine whether a suspected object is present in the area of ​​the soil spanned by the exploratory boreholes of the borehole arrangement, if the area of ​​the soil spanned by the exploratory boreholes of the borehole arrangement has not been categorized in the first category, in particular if the area of ​​the soil spanned by the exploratory boreholes of the borehole arrangement has been categorized in the second or third category.

[0032] It is particularly preferred that the bore arrangement has exactly three probe bores and that the probe bores of the bore arrangement have substantially equal distances to each other or that the distances differ from each other by no more than 10%, wherein preferably the connecting lines between the probe bore axes of the probe bores of the bore arrangement span an equilateral triangle in a horizontal plane.

[0033] It is particularly preferred that the algorithm be a random forest algorithm. A random forest algorithm is specifically an algorithm used in machine learning and can be described as an ensemble method employed in classification and regression procedures. When training such an algorithm, several decision trees with as little correlation as possible are preferably generated. Each decision tree is trained with a different, randomly selected sample of the training data. Additionally, each tree considers only a randomly chosen subset of all features at each node when distributing the objects from its sample. Subsequently, all trees are combined into an ensemble, the random forest.

[0034] The Random Forest algorithm is particularly advantageous for this categorization because, firstly, the reduction to key values ​​does not require a consistent data structure for the measurement data, and secondly, the internal structure prevents or minimizes "overfitting" in highly noisy measurement data.

[0035] It is particularly preferred that a database containing soil information (geodatabase) is provided, in which the positions of the existing exploratory boreholes of the borehole array and the associated measurement data are recorded. A display device shows the positions of the existing exploratory boreholes of the borehole array on a soil map, and the borehole array is marked, in particular color-coded, according to its categorization. Preferably, the database also contains project information and / or soil boundary conditions and / or aerial photographs and / or soil layer lists.

[0036] It is particularly preferred that the method be carried out for a plurality of exploratory boreholes distributed across a soil area to be investigated, wherein the plurality of exploratory boreholes are preferably assigned to a plurality of borehole arrangements. It is also particularly possible that one exploratory borehole is assigned to several borehole arrangements. This can be the case, in particular, with borehole arrangements arranged in close proximity.

[0037] It is particularly preferred that the method comprises: automated determination of a maximum borehole spacing between planned exploratory boreholes to be drilled into the ground or between planned exploratory boreholes and existing exploratory boreholes, wherein measurement data generated by at least one sensor in at least one existing exploratory borehole are analyzed, and a noise signal for the corresponding existing exploratory borehole is determined based on the measurement data, wherein the maximum borehole spacing is determined as a function of the noise signal, and wherein the noise signal is preferably determined by applying a bandpass filter to the generated measurement data. The unfiltered, bandpass-filtered, or high-pass filtered measurement data can be used for this determination. A lower and an upper limit line can be determined from the measurement data over a selected depth distance.The maximum borehole spacing can then be determined from the difference between the two limit lines, i.e., the lower and the upper limit line.

[0038] A maximum borehole spacing is understood to mean, in particular, a borehole spacing—that is, a distance between the axes of two exploratory boreholes—that is such that it is still possible to determine whether a suspected object might be present within the corresponding borehole configuration. The term "maximum borehole spacing" thus indicates the maximum permissible borehole spacing for the detection of suspected objects between the corresponding exploratory boreholes or within the area of ​​the ground spanned by the exploratory boreholes to remain possible. A maximum borehole spacing is preferably understood to be such a distance that categorization into the first or second category for the corresponding borehole configuration is still possible.

[0039] The noise signal is understood to be, in particular, a signal of unavoidable background noise that is recorded by the measurement. The noise signal preferably indicates the proportion of noise to the signal or the measurement. The signal-to-noise ratio is preferably a measure of the technical quality of the measurement data. Preferably, the maximum borehole spacing is determined as a function of the signal-to-noise ratio.

[0040] Preferably, the method comprises: adjusting, in particular automatically adjusting, a new borehole spacing between planned exploratory boreholes or between planned exploratory boreholes and existing exploratory boreholes depending on the automatically determined maximum borehole spacing, wherein the borehole spacing between the planned exploratory boreholes or between planned exploratory boreholes and existing exploratory boreholes is particularly preferably updated regularly, preferably at least daily, depending on the automatically determined maximum borehole spacings. Preferably, the borehole spacing between the planned exploratory boreholes or between planned exploratory boreholes and existing exploratory boreholes is updated at least hourly, particularly preferably at least every 30 minutes, and in particular in real time, i.e., immediately, depending on the automatically determined maximum borehole spacings.

[0041] Preferably, the new borehole spacing is adjusted several times during the soil investigation. Preferably, measurement data obtained from the most recently drilled borehole or boreholes are taken into account. This allows the maximum borehole spacing to be updated continuously and immediately after measurements are carried out, which is particularly advantageous. In this way, the maximum borehole spacing can be determined locally based on the measurement data, especially the signal-to-noise ratio, and subsequent boreholes can be drilled directly with such a locally adjusted maximum borehole spacing.

[0042] One advantage of such a method is that the borehole spacing determination can be automated and visually processed as soon as new measurement data is loaded. Preferably, a fixed bandpass filter is applied to the measurement data, from which the overall smoke signal of each exploratory borehole can be determined using various calculations of medians and mean values.

[0043] It is particularly preferred that the corresponding maximum borehole spacing is determined automatically and immediately, in particular within a maximum of 10 minutes, preferably a maximum of 5 minutes, particularly preferably a maximum of one minute, especially in real time, i.e. immediately, after generating the measurement data for the corresponding existing sounding borehole and / or after generating the noise signal for the corresponding existing sounding borehole.

[0044] It is particularly preferred that for each existing probe borehole, a detection radius is determined depending on the maximum borehole spacing assigned to the existing probe borehole, which defines a circle around the existing probe borehole in which effective detection of suspected objects is possible based on the measurement data generated in this existing probe borehole.

[0045] The detection radius refers in particular to a kind of effective radius that specifies the radius originating from and orthogonal to the probing bore axis of the probing bore, defining a circular area around the probing bore axis in which detection of suspected objects is possible.

[0046] It is particularly preferred that a database be provided by means of which the existing exploratory boreholes are linked, wherein the positions of the existing exploratory boreholes are displayed on a soil map by means of a display device, wherein the circle around the existing exploratory boreholes is marked, in particular by color. In this way, a visual representation of the soil coverage and, in particular, of any gaps that may still exist outside the detection radii surrounding the exploratory boreholes is obtained.

[0047] Preferably, the detection radius for the corresponding existing probe bore and the corresponding circle around the existing probe bore are determined automatically and immediately, in particular within a maximum of 10 minutes, preferably a maximum of 5 minutes, particularly preferably a maximum of one minute, in particular in real time, i.e. immediately, after the measurement data for the existing probe bore have been generated, and preferably displayed.

[0048] The detection radius is preferably displayed visually on the display device, both in color and in a size scaled relative to the detection radius. The calculation and visualization can also be performed for different depth ranges, i.e., for varying depths, particularly in relation to the depth along the borehole axes. This allows for the immediate identification of problematic areas and faster planning of further exploration procedures. Such a display is also particularly well-suited for discussing the exploration results with non-experts.

[0049] It is particularly preferred that the method comprises: creating further exploratory boreholes in the ground in an area where there is no coverage by circles defined by the detection radius for the corresponding existing exploratory boreholes.

[0050] Preferably, at least one detection radius of at least one existing sounding borehole is smaller compared to detection radii of existing sounding boreholes in other adjacent areas of the soil.

[0051] It is particularly preferred that the performance of at least one physical measurement, at least for some of the sounding bores, preferably for all of the sounding bores, comprises performing an electromagnetic measurement and a magnetic measurement, and preferably a radar measurement.

[0052] It is also possible and preferred that data generated by means of several physical measurements are evaluated in combination.

[0053] Preferably, at least one physical measurement is carried out for at least some exploratory boreholes, and preferably for all of them, using a passive method. Passive methods are based on measuring distortions of the Earth's magnetic field caused by magnetically active obstructions, such as unexploded ordnance, iron fragments, etc., in the subsurface. The magnetometers used for this purpose are highly sensitive measuring systems capable of detecting disturbances in the Earth's magnetic field caused by ferromagnetic minerals and objects. Magnetometers can be used in conjunction with recording and evaluation systems for borehole sounding when searching for suspected objects, especially munitions, at greater depths or in surface disturbances.

[0054] Preferably, at least one physical measurement is carried out for at least some exploratory boreholes, and preferably for all of the exploratory boreholes, using an active method. In active methods, an electromagnetic field is generated by a transmitter, the response of which from the subsurface is modified by electrically conductive structures such as metal bodies. These modifications are recorded and evaluated by a receiver. The pulse induction method that can be used in this context is based on the induction of eddy currents in an electrically conductive disturbance.

[0055] It is particularly preferred that a magnetic measurement is carried out for all sounding boreholes and an electromagnetic measurement is carried out for selected sounding boreholes, especially for those sounding boreholes where the area of ​​the soil spanned by the sounding boreholes of the borehole arrangement, to which the sounding boreholes belong, has not been categorized in the first category.

[0056] The performance of at least one physical measurement for at least some of the exploratory boreholes is preferably carried out by means of magnetic measurement, wherein preferably an electromagnetic measurement is additionally carried out for some selected or all exploratory boreholes.

[0057] Using two different measurement methods, namely magnetic measurement combined with electromagnetic measurement, has the advantage that suspected objects can be detected even better and more reliably. This applies particularly to unexploded bombs, which are magnetic and conductive.

[0058] It is particularly preferred that the measurement data generated by magnetic measurement and the measurement data generated by electromagnetic measurement are taken into account when categorizing the soil in the area of ​​the corresponding borehole arrangement and / or when automatically determining a maximum borehole spacing.

[0059] It is particularly preferred that a sounding borehole is digitally generated as a simulation and that simulated measurement data are generated for the sounding borehole generated as a simulation, depending on measurement data generated for existing real sounding boreholes that are located adjacent to the sounding borehole generated as a simulation.

[0060] Preferably, the method is used to plan exploratory boreholes. Preferably, additional boreholes in critical areas are simulated. The signal from the surrounding exploratory boreholes is preferably weighted to simulate the noise level of the planned borehole. A detection radius is derived from such a simulated signal, calculated according to an equation based on the distance between the exploratory boreholes. This allows for a preferably graphical estimate of how many additional exploratory boreholes are needed in a selected area to cover the ground surface. This is preferably done computationally and can be manually adjusted and / or modified based on the graphical representation. Preferably, an optimization algorithm is used to determine the most effective number and positions of boreholes to fill the previously uncovered areas.Preferably, manual adjustments are still possible afterwards.

[0061] Preferably, additional exploratory boreholes are planned and / or created based on the simulated measurement data. Particularly preferably, these additional boreholes are planned and / or created in areas of the soil where the detection radii of adjacent boreholes do not overlap.

[0062] It is particularly preferred that a suspected object is digitally generated as a simulation and that it is then checked whether the existing real exploratory boreholes arranged adjacent to the suspected object generated as a simulation are close enough to the suspected object generated as a simulation to be able to detect a potential suspected object at the position of the suspected object generated as a simulation.

[0063] Preferably, depending on the result of this test, additional exploratory boreholes are planned and / or created that are located closer to the suspected object generated as a simulation than the existing exploratory boreholes.

[0064] According to a further aspect of the invention, the aforementioned problem is solved by a system for detecting suspected objects, in particular munitions, in soil, wherein the system is preferably configured to carry out a method as described herein, the system comprising a device for creating exploratory boreholes and for performing measurements in exploratory boreholes, wherein the device is configured to perform the following steps: creating at least three, in particular exactly three, spaced-apart exploratory boreholes in soil, wherein the at least three exploratory boreholes form a borehole arrangement, preferably a borehole triplet, wherein preferably the at least three exploratory boreholes of the borehole arrangement each have a exploratory borehole axis that is arranged parallel to each other and / or vertically aligned;Performing at least one physical measurement within each of the at least three sounding boreholes of the borehole array using at least one sensor; generating measurement data based on the at least one physical measurement performed for each of the at least three sounding boreholes of the borehole array; a data processing system trained to perform the following steps: analyzing the measurement data associated with the borehole array using an algorithm trained on a training dataset to categorize an area of ​​the ground spanned by the sounding boreholes of the borehole array;Categorizing the soil in the area of ​​the borehole array, depending on the measurement data assigned to the borehole array, using the algorithm into one of at least two categories, wherein the at least two categories comprise: a first category according to which the measurement data indicate that no suspected object is present in the area of ​​the soil spanned by the exploratory boreholes of the borehole array; a second category according to which the measurement data indicate that a suspected object is present in the area of ​​the soil spanned by the exploratory boreholes of the borehole array.

[0065] The system is preferably designed to carry out further investigations to determine whether a suspected object is present in the area of ​​the soil spanned by the probe boreholes of the borehole arrangement, if the soil in the area of ​​the borehole arrangement has not been categorized into the first category, in particular if the soil in the area of ​​the borehole arrangement has been categorized into the second category.

[0066] According to another aspect of the invention, the aforementioned problem is solved by using a system as described here for detecting suspected objects, in particular munitions, in soil.

[0067] For the advantages, design variants and design details of the various aspects of the solutions described here and their respective possible further developments, reference is also made to the description of the corresponding features, details and advantages of the other aspects and their further developments.

[0068] Preferred embodiments are explained by way of example with reference to the accompanying figures. The drawings are not necessarily to scale. In the figures, identical or essentially functionally equivalent or similar elements are designated with the same reference numerals. They show: Fig. 1: a schematic representation of a procedure for detecting suspected objects, especially munitions, in soil; Fig. 2: a schematic side view of a device for creating exploratory boreholes in soil and for detecting suspected objects; Fig. 3: a schematic representation of a borehole arrangement in the form of a borehole triplet; Fig. 4: a first exemplary representation of a display device; Fig. 5: a second exemplary representation of a display device.

[0069] Fig. Figure 1 shows a schematic representation of a procedure 100 for detecting suspected objects, especially munitions, in soil. The procedure comprises the following steps:

[0070] In step 110, producing at least three, in particular exactly three, spaced-apart sounding boreholes in a soil, wherein the at least three sounding boreholes form a borehole arrangement, preferably a borehole triple, wherein preferably the at least three sounding boreholes of the borehole arrangement each have a sounding borehole axis S which are arranged parallel to each other and / or are vertically aligned.

[0071] In step 120, perform at least one physical measurement within each of the at least three probe bores of the bore arrangement using at least one sensor.

[0072] In step 130, generating measurement data based on the at least one physical measurement performed for each of the at least three probe boreholes of the borehole arrangement.

[0073] In step 140, the measurement data associated with the borehole array are analyzed using an algorithm trained on the basis of a training data set to categorize an area of ​​the ground spanned by the probe boreholes of the borehole array.

[0074] In step 150, categorizing the area of ​​the ground spanned by the probe boreholes of the borehole arrangement, depending on the measurement data assigned to the borehole arrangement, using the algorithm into one of at least two categories, wherein the at least two categories comprise: a first category according to which the measurement data indicate that no suspected object is present in the area of ​​the ground spanned by the probe boreholes of the borehole arrangement; a second category according to which the measurement data indicate that a suspected object is present in the area of ​​the ground spanned by the probe boreholes of the borehole arrangement.

[0075] In step 160, conduct further investigations to determine whether a suspected object is present in the area of ​​the drilling arrangement, in particular in the area of ​​the ground spanned by the sounding boreholes of the drilling arrangement, if the area of ​​the ground spanned by the sounding boreholes of the drilling arrangement has not been categorized into the first category, in particular if the area of ​​the ground spanned by the sounding boreholes of the drilling arrangement has been categorized into the second category.

[0076] Preferably, the method also includes the following steps, which can be carried out in particular after step 130: In one step 200, automated determination of a maximum borehole spacing between planned exploratory boreholes to be drilled into the ground or between planned exploratory boreholes and existing exploratory boreholes, wherein measurement data generated by means of at least one sensor in at least one existing exploratory borehole are analyzed and a noise signal for the corresponding existing exploratory borehole is determined on the basis of the measurement data, wherein the maximum borehole spacing is determined as a function of the noise signal, wherein the noise signal is preferably determined by applying a bandpass filter to the generated measurement data.

[0077] In step 210, adjustment, in particular automated adjustment of a new borehole spacing between planned exploratory boreholes or between planned exploratory boreholes and existing exploratory boreholes depending on the automatically determined maximum borehole spacing, wherein the borehole spacing between the planned exploratory boreholes or between planned exploratory boreholes and existing exploratory boreholes is particularly preferably updated regularly, preferably at least daily, depending on the automatically determined maximum borehole spacings.

[0078] Fig. Figure 2 shows a System 500 for detecting suspected objects, in particular munitions, in soil. The System 500 includes a device 1 for creating exploratory boreholes SB in soil B and for detecting suspected objects, in particular munitions, in the soil B.

[0079] The device 1 comprises a carrier unit 10. In the example shown, the carrier unit 10 is designed as a construction machine in the form of a crawler excavator. The carrier unit 10 has a drive 15 and a means of locomotion 16. Furthermore, the carrier unit 10 has a boom 11, which is connected to a drill chuck 20. The boom 11 is rotatable about the boom axis of rotation 1100, which is arranged vertically, i.e., in the vertical direction V. This also allows the drill chuck 20 to rotate about the boom axis of rotation 1100.

[0080] The device 1 further comprises a drill 30 having a vertically oriented longitudinal axis 300. The drill 30 is connected to the drill holder 20 and is axially movable on the drill holder 20 in the direction of the longitudinal axis 300 and is also rotatably mounted about the longitudinal axis 300. A drill motor 40 is provided, which is connected to the drill 30 and is configured to drive the drill 30 to produce a test bore SB with a vertically arranged test bore axis S in the ground B using the drill 30.

[0081] The device 1 further comprises a probe rod (in Fig. 2 not visible), which can be designed and arranged as shown in the figures below. A sensor 90 is arranged on the probe rod, which can be inserted into the probe bore SB in order to perform at least one physical measurement within the probe bore SB, so that measurement data can then be generated based on these measurements.

[0082] Furthermore, a database 700, which can also be referred to as a geodatabase, is provided, through which the existing exploratory boreholes are linked. The positions of the existing exploratory boreholes within the borehole array and the generated measurement data are recorded in the database 700. Using a display unit 600, the positions of the existing exploratory boreholes can be displayed on a soil map, whereby the borehole array can be marked, in particular color-coded, depending on the categorization performed for the borehole array.

[0083] Fig. Figure 3 shows a schematic representation of a borehole arrangement BA in the form of a triple borehole. Three exploratory boreholes SB are provided, which, in plan view, form an equilateral triangle with their connecting lines corresponding to the borehole spacing SBA, also known as the exploratory borehole spacing. The detection radius DR is specified for the exploratory borehole SB shown below. The detection radius DR defines the circle K, which defines the area around the exploratory borehole SB in which the detection of suspected objects is possible. The position of a suspected object, identified using the measurement data generated in the exploratory boreholes SB, is indicated by VO.

[0084] Fig. Figure 4 shows a first exemplary representation of a display device 600. Using the display device 600, the positions of the existing exploratory boreholes can be displayed on a soil map, with the borehole arrangements being marked, in particular color-coded, depending on the categorization performed for the borehole arrangements. The detection radii are shown in Fig. The data is displayed both in color and with a size corresponding to a scaled size of the determined detection radius. The calculation and visualization can be performed for various depth ranges. Green indicates exploratory boreholes with a sufficiently large detection radius to cover the corresponding area. Yellow, orange, and red, respectively, indicate exploratory boreholes with relatively small detection radii. In these areas, the detection radii are not large enough to cover the corresponding area with the circles defined by the detection radii. Using this view, problematic areas can be immediately identified, and further investigations can be carried out in these problematic areas.

[0085] Fig. Figure 5 shows a second exemplary representation of a display unit 600. The one in Fig. The view shown in Figure 5 is used for planning new exploratory boreholes. Additional boreholes in critical areas can be simulated. The signal from measurements taken in the surrounding boreholes is weighted to simulate the noise level of the planned borehole. These weights can also be adjusted manually. The simulated signal yields an influence radius, which is calculated using a formula for determining borehole spacing. This allows for a preliminary, and then graphical, estimation of how many additional boreholes are needed in a selected area to cover the area as determined by the borehole spacing. This is also possible for different depth ranges. Using this method for planning additional boreholes, it is also possible to automatically fill a predefined area with planned boreholes.For example, an iterative approach with different overlap limits can be used to fill the selected area with the fewest possible holes and cover it almost completely, as shown in the view in . Fig. 5 shown. Reference symbol list 1 Device 10 Carrier device 11 outriggers 15 Drive 16 means of transport 20 drill chuck arrangement 30 drill bits 40 drill motor 90 Sensor 100 methods for detecting suspected objects, especially munitions, in soil 110-210 process steps 300 drill longitudinal axis 500 System for detecting suspicious objects, especially munitions, in soil 600 Display unit 700 database 1100 Boom pivot axis B Floor BA drilling arrangement DR detection radius K Circle S Exploratory drilling axis SB exploratory drilling SBA borehole spacing, exploratory borehole spacing MBA Maximum borehole spacing V vertical direction VO Suspected Object

Claims

Method (100) for detecting suspected objects, in particular munitions, in soil (B), the method comprising: - creating (110) at least three, in particular exactly three, spaced-apart probe boreholes (SB) in soil (B), wherein the at least three probe boreholes (SB) form a borehole arrangement (BA), preferably a borehole triple, wherein preferably the at least three probe boreholes (SB) of the borehole arrangement (BA) each have a probe borehole axis (S) which are arranged parallel to each other and / or are vertically aligned; - performing (120) at least one physical measurement within each of the at least three probe boreholes (SB) of the borehole arrangement (BA) by means of at least one sensor (90); - generating (130) measurement data based on the at least one physical measurement performed for each of the at least three probe boreholes (SB) of the borehole arrangement (BA);- Analyzing (140) the measurement data associated with the borehole array (BA) using an algorithm trained on the basis of a training dataset to categorize an area of ​​the ground (B) spanned by the sounding boreholes (SB) of the borehole array (BA); - Categorizing (150) the area of ​​the ground (B) spanned by the sounding boreholes (SB) of the borehole array (BA) in relation to the measurement data associated with the borehole array (BA) using the algorithm into one of at least two categories, wherein the at least two categories comprise: ◯ a first category according to which the measurement data indicate that no suspected object is present in the area of ​​the ground spanned by the sounding boreholes of the borehole array; ◯ a second category according to which the measurement data indicate that a suspected object is present in the area of ​​the ground spanned by the sounding boreholes of the borehole array;- Carry out (160) further investigations to determine whether a suspected object (VO) is present in the area of ​​the drilling arrangement (BA), in particular in the area of ​​the soil (B) spanned by the exploratory boreholes (SB) of the drilling arrangement (BA), if the area of ​​the soil (B) spanned by the exploratory boreholes (SB) of the drilling arrangement (BA) has not been categorized in the first category, in particular if the area of ​​the soil (B) spanned by the exploratory boreholes (SB) of the drilling arrangement (BA) has been categorized in the second category.; The method of claim 1, wherein the categorization (150) is carried out by means of the algorithm into one of at least three categories, the at least three categories comprising: the first category, the second category, and a third category, according to which, in particular due to interference signals, categorization into the first category or into the second category is not possible, wherein preferably the carrying out (160) of further investigations to determine whether a suspected object (VO) is present in the area of ​​the soil (B) spanned by the sounding boreholes (SB) of the borehole arrangement (BA) is carried out if the area of ​​the soil (B) spanned by the sounding boreholes (SB) of the borehole arrangement (BA) has not been categorized into the first category, in particular if the area of ​​the soil (B) spanned by the sounding boreholes (SB) of the borehole arrangement (BA) has been categorized into the second category or into the third category. Method according to one of the preceding claims, wherein the bore arrangement (BA) has exactly three probe bores (SB) and the probe bores (SB) of the bore arrangement (BA) each have substantially equal distances to each other or the distances differ from each other by no more than 10%, wherein preferably the connecting lines between the probe bore axes (S) of the probe bores (SB) of the bore arrangement (BA) span an equilateral triangle in a horizontal plane. Method according to one of the preceding claims, wherein the algorithm is a random forest algorithm. Method according to one of the preceding claims, wherein a database containing information on the soil (B) (geodatabase) is provided, in which the positions of the existing exploratory boreholes (SB) of the borehole arrangement (BA) and the generated measurement data are recorded, wherein the positions of the existing exploratory boreholes (SB) of the borehole arrangement (BA) are displayed on a soil map by means of a display device (600), wherein the borehole arrangement (BA) is marked, in particular color-coded, depending on the categorization carried out for the borehole arrangement (BA). Method according to one of the preceding claims, wherein the method is carried out for a plurality of sounding boreholes (SB) distributed over a soil area to be investigated, wherein the plurality of sounding boreholes (SB) are preferably assigned to a plurality of borehole arrangements (BA). A method according to one of the preceding claims, the method comprising: - automated determination (200) of a maximum borehole spacing (MBA) between planned exploratory boreholes (SB) to be drilled into the (B) ground or between planned exploratory boreholes (SB) and existing exploratory boreholes (SB), wherein measurement data generated by means of at least one sensor (90) in at least one existing exploratory borehole (SB) are analyzed and a noise signal for the corresponding existing exploratory borehole (SB) is determined on the basis of the measurement data, wherein the maximum borehole spacing (MBA) is determined as a function of the noise signal, wherein the noise signal is preferably determined by applying a bandpass filter to the generated measurement data, - and preferably adapting (210),In particular, automated adjustment of a new borehole spacing (SPS) between planned exploratory boreholes (EB) or between planned exploratory boreholes (EB) and existing exploratory boreholes (EB) depending on the automatically determined maximum borehole spacing (MPS), wherein the borehole spacing (SPS) between the planned exploratory boreholes (EB) or between planned exploratory boreholes (EB) and existing exploratory boreholes (EB) is preferably updated regularly, preferably at least daily, depending on the automatically determined maximum borehole spacings (MPS). Method according to claim 7, wherein the corresponding maximum borehole spacing (MBA) is determined automatically and immediately, in particular within a maximum of 10 minutes, preferably a maximum of 5 minutes, particularly preferably a maximum of one minute, in particular in real time, i.e. immediately, after generating the measurement data for the corresponding existing sounding borehole (SB) and / or after generating the noise signal for the corresponding existing sounding borehole (SB). Method according to one of claims 7 - 8, wherein for each existing probe borehole (SB) a detection radius (DR) is determined depending on the maximum borehole spacing (MBA) assigned to the existing probe borehole (SB), which defines a circle (K) around the existing probe borehole (SB) in which effective detection of suspected objects (VO) is possible based on the measurement data generated in this existing probe borehole (SB). A method according to claim 9, wherein a database (700) is provided by means of which the existing exploratory boreholes (EB) are linked, wherein the positions of the existing exploratory boreholes (EB) are displayed on a ground map by means of a display device (600), wherein the circle (K) around the existing exploratory boreholes (EB) is marked, in particular marked in color, and displayed, wherein preferably the detection radius (DR) for the corresponding existing exploratory borehole (EB) and the corresponding circle (K) around the existing exploratory borehole (EB) are determined automatically and immediately, in particular within a maximum of 10 minutes, preferably a maximum of 5 minutes, particularly preferably a maximum of one minute, in particular in real time, i.e. immediately, after the measurement data for the existing exploratory borehole (EB) have been generated, and preferably displayed. Method according to claim 10, the method comprising: - generating (110) further sounding boreholes (SB) in the ground (B) in an area in which there is no coverage by circles (K) defined by the detection radius (DR) for the corresponding existing sounding boreholes (SB). A method according to one of the preceding claims, wherein performing at least one physical measurement at least for some of the sounding boreholes (SB), preferably for all of the sounding boreholes (SB), comprises performing an electromagnetic measurement and a magnetic measurement, and preferably a radar measurement, and / or wherein a magnetic measurement is performed for all sounding boreholes (SB) and an electromagnetic measurement is additionally performed for selected sounding boreholes (SB), in particular for those sounding boreholes (SB) where the area of ​​the ground (B) spanned by the sounding boreholes (SB) of the borehole arrangement (BA), to which the sounding boreholes (SB) belong, has not been categorized in the first category. Method according to claim 12, wherein the measurement data generated by magnetic measurement and the measurement data generated by electromagnetic measurement are taken into account when categorizing (150) the soil in the area of ​​the corresponding borehole arrangement (BA) and / or when automatically determining (200) a maximum borehole spacing (MBA). Method according to one of the preceding claims, wherein a sounding borehole (SB) is digitally generated as a simulation and simulated measurement data are generated for the sounding borehole (SB) generated as a simulation depending on measurement data generated for existing real sounding boreholes (SB) that are arranged adjacent to the sounding borehole (SB) generated as a simulation, wherein preferably additional sounding boreholes (SB) are planned and / or generated depending on the simulated measurement data. A method according to one of the preceding claims, wherein a suspected object (VO) is digitally generated as a simulation and it is subsequently checked whether the existing real exploratory boreholes (SB) arranged adjacent to the suspected object (VO) generated as a simulation are close enough to the suspected object (VO) generated as a simulation to be able to detect a potential suspected object (VO) at the position of the suspected object (VO) generated as a simulation, wherein preferably, depending on the result of this check, additional exploratory boreholes (SB) which are arranged closer to the suspected object (VO) generated as a simulation than the already existing exploratory boreholes (SB) are planned and / or generated. System (500) for detecting suspected objects, in particular munitions, in soil (B), wherein the system is preferably configured to perform a method (100) according to one of the preceding claims, the system comprising: a device (1) for creating exploratory boreholes (SB) and for performing measurements in exploratory boreholes (SB), wherein the device (1) is configured to perform the following steps: o Creating (110) at least three, in particular exactly three, spaced-apart exploratory boreholes (SB) in soil (B), wherein the at least three exploratory boreholes (SB) form a borehole arrangement (BA), preferably a borehole triple, wherein preferably the at least three exploratory boreholes (SB) of the borehole arrangement (BA) each have a exploratory borehole axis (S) which are arranged parallel to each other and / or are vertically aligned;o Performing (120) at least one physical measurement within each of the at least three sounding boreholes (SB) of the borehole arrangement (BA) using at least one sensor; o Generating (130) measurement data based on the at least one physical measurement performed for each of the at least three sounding boreholes (SB) of the borehole arrangement (BA); - a data processing system configured to perform the following steps: o Analyzing (140) the measurement data associated with the borehole arrangement (BA) using an algorithm trained on the basis of a training dataset to categorize an area of ​​the ground spanned by the sounding boreholes (SB) of the borehole arrangement (BA);o Categorizing (150) the soil in the area of ​​the borehole arrangement (BA) depending on the measurement data associated with the borehole arrangement (BA) using the algorithm into one of at least two categories, wherein the at least two categories comprise: ▪ a first category according to which the measurement data indicate that no suspected object (VO) is present in the area of ​​the soil (B) spanned by the sounding boreholes (SB) of the borehole arrangement (BA); ▪ a second category according to which the measurement data indicate that a suspected object (VO) is present in the area of ​​the soil (B) spanned by the sounding boreholes (SB) of the borehole arrangement (BA);wherein the system is preferably designed to carry out further investigations to determine whether a suspected object (VO) is present in the area of ​​the soil (B) spanned by the exploratory boreholes (SB) of the borehole arrangement (BA), if the soil (B) in the area of ​​the borehole arrangement (BA) has not been categorized into the first category, in particular if the soil in the area of ​​the borehole arrangement has been categorized into the second category. Use of a system according to the preceding claim for detecting suspected objects (VO), in particular munitions, in a soil (B).

Images