Methods for detecting underground structures

Abstract

The present invention relates to a method and apparatus for detecting underground structures. The method comprises the steps of sending electromagnetic waves into the structure, receiving echoes of the electromagnetic waves from the structure, and processing the echoes to derive internal characteristics of the structure. The step of sending electromagnetic waves into the structure comprises subsequently sending a plurality of electromagnetic detection signals having different spectra into the structure. Each detection signal comprises at least two non-zero spectral components. The step of receiving the echoes comprises receiving an echo signal for each detection signal. The step of processing the echoes comprises determining at least one amplitude and phase for each echo signal.

Description

Technical Field

[0001] The present invention relates to a method and apparatus for detecting underground structures using electromagnetic waves. Background Art

[0002] Electromagnetic waves are commonly used to detect underground structures, such as man-made structures or beneath the ground. Ground penetrating radar (GPR) is a technique commonly used in geophysical surveys or non-destructive testing of, for example, concrete structures, which employs electromagnetic waves typically in the range of 10 MHz to 2.6 GHz.

[0003] A known method for GPR data acquisition is stepped frequency continuous wave (SFCW). Instead of sending a single broadband pulse, SFCW uses a continuous wave probe signal of a certain frequency that changes over time or between subsequent probe signals. Examples of such methods and suitable equipment are described in WO2018 / 161183A1.

[0004] The SFCW method is advantageous in terms of data quality and therefore in terms of the quality of the resulting image of the subsurface structure, in particular in terms of signal-to-noise ratio (SNR) and resolution. On the other hand, such methods are limited in acquisition speed, which means that GPR measurements of a given resolution (e.g., 2 cm) can only be obtained if the movement speed of the GPR device relative to the subsurface does not exceed a maximum acquisition speed, which is typically of the order of, for example, 10 to 20 km / h. Summary of the invention

[0005] Therefore, a general object of the present invention is to provide a method and an apparatus for detecting underground structures which allow faster data acquisition, in particular a higher acquisition speed of the apparatus for detecting underground structures, while maintaining a high resolution of the acquired data.

[0006] Features described below with respect to the method are intended to also apply to the device, and vice versa.In addition, the described features are intended to be disclosed independently of each other and in combination where reasonable.

[0007] Method for detecting underground structures

[0008] In order to achieve the above and further objects of the present invention which will become more apparent in the following description, a method for detecting underground structures (e.g., man-made structures or below the ground) comprises the following steps:

[0009] - Sending electromagnetic waves into the structure: The electromagnetic waves may be polarized, for example linearly polarized. The step of sending electromagnetic waves into the structure is performed in particular by means of an antenna.

[0010] - Receiving an echo of an electromagnetic wave from a structure: Advantageously, a portion of the wave sent into the structure is reflected by an internal feature of the structure characterized by a change in an electrical property, such as a change in the dielectric constant. The reflected portion of the wave can be received from the structure as an echo (i.e., a reflected electromagnetic wave).

[0011] -Processing the echoes to derive internal features of the structure: for example, the internal features may be the location and / or gradient of internal boundaries within the structure, characterization of scatterers or reflectors, changes in electrical properties (e.g., dielectric constant), specific locations in the structure, etc.

[0012] The step of sending electromagnetic waves into the structure comprises subsequently sending a plurality of electromagnetic detection signals having different spectra into the structure. In particular, the spectrum of the detection signal is given by its Fourier transform. Each detection signal comprises at least two non-zero spectral components. In particular, this is different from conventional SFCW, in which each detection signal has only one spectral component, meaning that only one frequency is sent at a time.

[0013] In an embodiment, each detection signal comprises more than two non-zero spectral components, in particular 3 to 20, in particular 5 to 15, in particular 10 non-zero spectral components. In an extreme case, each detection signal may comprise noise, such as white noise, on a given spectrum, which advantageously has a width of at least 1000 MHz.

[0014] The step of receiving the echo includes receiving an echo signal for each detection signal. In particular, the echo signal is received subsequently.

[0015] The step of processing the echoes comprises determining at least one amplitude and a phase for each echo signal, advantageously one for a plurality of spectral components, in particular for each non-zero spectral component.

[0016] The described method for detecting underground structures has the advantage that a plurality of spectral components, in particular at least two, are sent concurrently (i.e. simultaneously) into the structure. Compared to conventional SFCW, this method can increase the data acquisition speed by a factor equal to the number of spectral components in each detection signal. This means that the corresponding device for detecting underground structures can be moved at a higher speed, in particular at a higher speed equal to a factor equal to the number of spectral components in each detection signal, while maintaining the same resolution. This in turn facilitates mounting the device, for example on a car or on an aerial drone, and performing data acquisition over large areas and / or inaccessible locations.

[0017] On the other hand, the described method also presents challenges and has further advantages, which will be described in conjunction with the following advantageous embodiments.

[0018] Orthogonal spectral components

[0019] In an advantageous embodiment, the spectral components are concurrent subcarrier signals with different subcarrier frequencies. In particular, the spectral components are located in non-overlapping frequency bands. In the field of telecommunications, a similar approach is referred to as frequency division multiplexing (FDM) in a completely different context. Non-overlapping frequency bands allow covering a wide spectrum range even with low-power transmitters.

[0020] In a further advantageous embodiment, at least two spectral components are orthogonal to each other. "Orthogonal" means in particular that they satisfy an orthogonality condition, usually defined as their inner product, or equivalently as the integral of their product over a time interval, being zero. In particular, if the time integral ∫f(t)g(t)dt of the product of a first subcarrier signal f(t) and a second subcarrier signal g(t) is smaller than a finite threshold, e.g. much smaller than √[∫f(t)f(t)dt·∫g(t)g(t)dt], in particular smaller than 10% of this expression, then the two signals are said to be "orthogonal" in the present context. In particular, the time integral is calculated over a time interval representing the periodicity of f(t) and / or g(t). Examples of orthogonal subcarrier signals are functions of the form sin(mx) and sin(nx), where m and n are unequal positive integers. In the field of telecommunications, similar applications are known as orthogonal frequency division multiplexing (OFDM).

[0021] In particular, adjacent subcarrier frequencies differ from each other by a subcarrier spacing, which may be a regular subcarrier spacing, i.e. the same for all pairs of adjacent subcarrier frequencies. Advantageously, the subcarrier spacing satisfies an orthogonality condition, e.g. in that the subcarrier frequency is an integer multiple of the subcarrier spacing. In particular, the subcarrier spacing may be between 1 and 10 MHz, in particular about 4 MHz.

[0022] In an embodiment, the spectrum of the detection signal lies between 10 and 8000 MHz, in particular between 40 and 3440 MHz. Advantageously, the total spectrum of a plurality of subsequently transmitted detection signals has a width of at least 500 MHz, in particular at least 1000 MHz, more in particular at least 2000 MHz. With such a total spectrum, the method is able to achieve a high resolution, for example 2 cm or less, while achieving a large penetration depth, for example 5 m or more, depending on the electrical characteristics of the underground structure.

[0023] Non-zero initial phase shift

[0024] Now, sending a detection signal comprising multiple orthogonal subcarrier signals without further measures results in a high peak transmit power. This is due to the fact that at some point, the peaks of the orthogonal functions with zero initial phase shift will constructively add up, generating a high cumulative signal. The "initial phase shift" can be defined as the phase difference between the zero phase of the lower frequency function and the closest zero phase of the higher frequency function. In the device for detecting underground structures, the front end (i.e., the antenna plus analog components for processing the received signal) as well as the detection signal generator for generating the detection signal and the echo signal processor for processing the echo signal must be adjusted to handle the peak transmit power. Due to signal-to-noise ratio and low intermodulation reasons, it is not desirable to have a high peak transmit power that is much larger than the average transmit power.

[0025] Therefore, in an advantageous embodiment, at least two of the spectral components have a non-zero initial phase shift. In particular, the initial phase shift of at least two spectral components is such that the maximum amplitude of the detection signal is smaller than the maximum amplitude of a hypothetical detection signal having the same spectral components but with zero initial phase shift of the spectral components, in particular at least 10%, at least 25% or at least 50% smaller. This allows the peak transmit power to be kept at a low level, in particular not much larger than the average transmit power, for example only 2 or 3 times the average transmit power. Thus, a good signal-to-noise ratio and low intermodulation can be achieved.

[0026] In practice, for a given number of subcarrier signals, the initial phase shifts that satisfy the above conditions can be obtained by empirical methods, for example, in numerical simulations: the subcarrier signals are added with random initial phase shifts to generate test detection signals. The ratio of the maximum value of the test detection signal to the average value of the test detection signal is then derived. These two steps are iterated with different random initial phase shifts to form a set of test detection signals. For practical applications when detecting underground structures, the test detection signal with the smallest ratio is selected. In this method, the initial phase shifts can be obtained from a random number generator, but then they are not changed in the application. In particular, for each detection signal, the initial phase shifts of at least two spectral components can be constant, that is, the same.

[0027] Generate a detection signal

[0028] In an advantageous embodiment, the method additionally comprises the step of generating a detection signal by frequency converting at least two modulation signals of different modulation frequencies with at least one carrier signal. In particular, the step of generating the detection signal comprises changing the carrier frequency of the carrier signal between subsequent detection signals. In this way, the modulation frequency can be the same for all detection signals.

[0029] Advantageously, the steps of generating the detection signal and / or processing the echo signal are performed partially in the digital domain and partially in the analog domain, thereby taking advantage of the benefits of both domains, for example, precise control of the signal using existing digital data processing routines and the simplicity and speed of analog components. In an embodiment, the step of generating the detection signal comprises generating at least two digital modulation signals. Moreover, the step of generating the detection signal may comprise generating at least one analog carrier signal. According to the above description, at least two of the digital modulation signals advantageously have a non-zero initial phase shift.

[0030] Typical modulation frequencies are below 100 MHz, in particular in the range between 40 and 80 MHz. Advantageously, the modulation frequency differs according to the subcarrier spacing. In addition, the carrier frequency of the carrier signal may be at least 100 MHz.

[0031] In an embodiment, the step of generating the detection signal includes obtaining at least two analog modulation signals by performing digital-to-analog conversion on at least two digital modulation signals, and mixing the at least two analog modulation signals with at least one mixing signal from at least one analog oscillator by means of at least one analog mixer. In addition, the step of generating the detection signal may include sequentially mixing the analog modulation signals with mixing signals from a plurality of analog oscillators by means of a plurality of analog mixers.

[0032] Advantageously, the step of generating the detection signal comprises applying up-conversion to the analog modulated signal, and in particular up-conversion and subsequent down-conversion. Direct up-conversion usually requires an image rejection mixer, which brings some limitations, such as limited image rejection, local oscillator (LO) leakage and complex filter stages. Firstly up-converting the signal allows the image and LO leakage to be attenuated by the filter. Since the LO in the down-conversion stage is always higher than the desired output signal, the leakage of the LO can be ignored. In this way, complex filter stages can be avoided.

[0033] In a further embodiment, the step of generating a digital modulation signal comprises switching off a given modulation signal. This can still be performed in the digital domain, for example, by applying a switching stage. Switching off a given modulation signal changes the amount of modulation frequency in the corresponding detection signal. This in turn changes the peak voltage. A dynamic correction can then be applied to the modulation signal in order to keep the peak voltage constant, in particular at a desired level.

[0034] Additionally, it may be useful for the first of the detection signals to include a different number of non-zero spectral components than the second of the detection signals. For example, the higher frequency detection signal may contain fewer, e.g. half, spectral components than the lower frequency detection signal. In this case, the subcarrier spacing of the higher frequency detection signal may be greater than, e.g. twice, the subcarrier spacing of the lower frequency detection signal. This is particularly reasonable in order to minimize the data that needs to be processed and ultimately stored and transmitted. By enlarging the subcarrier spacing in the higher frequency detection signal, not much is lost in terms of penetration depth, whereas for high penetration depths, a smaller subcarrier spacing in the lower frequency detection signal may be required. Again, deletion of specific subcarrier signals may still be performed on the corresponding modulated signal in the digital domain.

[0035] In an embodiment, the total duration of each detection signal does not exceed 2 μs. Such a short duration of each detection signal prevents excessive electromagnetic wave energy from being sent into the environment, thereby possibly degrading telecommunication signals or affecting living things.

[0036] In addition, the envelope for each detection signal can have a rise time and a fall time, both of which have a non-zero duration, in particular both greater than 0.5 μs and / or both greater than 10% of the duration of each detection signal. Compared to a "hard window", in particular a hard window consisting of two step functions, the advantage of such a "smoothed" detection signal is the avoidance of ringing effects in the step of processing the echo. For that reason, the envelope can be shaped in particular as a Gaussian function or other known smooth window function.

[0037] Processing echo signals

[0038] Similar to the step of generating the detection signal, the step of processing the echo signal is also advantageously performed partly in the analog domain and partly in the digital domain, thereby combining the advantages of both as described above.

[0039] In an embodiment, the step of processing the echo signal comprises mixing the echo signal with at least one mixing signal from at least one analog oscillator by means of at least one analog mixer to obtain a mixed echo signal. In particular, the step of processing the echo signal may comprise sequentially mixing the echo signal with mixing signals from a plurality of analog oscillators by means of a plurality of analog mixers. It may be useful that the step of processing the echo signal comprises applying a down conversion (in particular an up conversion and a subsequent down conversion) to the echo signal.

[0040] Additionally, the step of processing the echo signal may include analog-to-digital converting the mixed echo signal into a digital echo signal.

[0041] Similar to above, the step of processing the echo signal advantageously comprises correcting the digital echo signal for an initial phase shift of at least two spectral components. In particular, the initial phase shift is known for each detection signal and thus for each corresponding echo signal and can therefore be corrected.

[0042] Furthermore, it may be desirable that the step of processing the echo signal comprises obtaining a serialized echo signal by serializing at least two non-zero spectral components, in particular using a digital echo signal. "Serialization" specifically refers to stringing together the non-zero spectral components in the echo signal (e.g., a modulation signal found from the echo signal) one after another in time. In this way, the serialized echo signal is similar to a signal obtained by a conventional SFCW method. This enables further processing using known data processing methods, for example, to assemble and process an A scan or apply a migration algorithm.

[0043] Finally, the step of processing the echo may include deriving at least one characteristic of the underground structure from an amplitude difference and / or a phase difference between the echo signal and the detection signal. The amplitude difference and / or the phase difference is directly related to, and in particular caused by, the material properties (in particular electrical properties) and material distribution encountered by the electromagnetic wave on its path of travel from the antenna into the structure and back.

[0044] Equipment for detecting underground structures

[0045] A second aspect of the present invention relates to a device for detecting underground structures. The device is configured to perform the above method and comprises:

[0046] -Antenna: In one embodiment, the same antenna can be used to send the detection signal and to receive the echo signal. In another embodiment, the antenna may include a first antenna configured to send the detection signal into the underground structure, and a second antenna configured to receive the echo signal received from the underground structure. In both embodiments, the antenna or the first and second antennas, respectively, include a lower frequency sub-antenna (e.g., with a center frequency of 350 MHz) and a higher frequency sub-antenna (e.g., with a center frequency of 1500 MHz). Such sub-antennas have the advantage of more efficiently transmitting and receiving the detection signal and the echo signal, respectively, at higher and lower frequencies, especially near their center frequencies.

[0047] -A detection signal generator for generating an electromagnetic detection signal to be sent by an antenna into an underground structure: In a preferred embodiment, the detection signal generator includes a digital signal generator configured to generate at least two digital modulation signals, a digital-to-analog converter configured to convert the digital modulation signal into an analog modulation signal, at least one analog signal generator configured to generate at least one carrier signal, and at least one analog mixer configured to generate a detection signal by mixing at least two analog modulation signals with at least one carrier signal.

[0048] -An echo signal processor for processing an echo signal received by an antenna from an underground structure: In a preferred embodiment, the echo signal processor comprises at least one further analog mixer configured to generate a mixed echo signal by mixing the echo signal with at least one carrier signal from at least one analog signal generator, an analog-to-digital converter configured to convert the mixed echo signal into a digital echo signal, and a digital processor configured to derive at least two non-zero spectral components from the digital echo signal.

[0049] - A control unit for controlling the detection signal generator and the echo signal processor: in particular, the control unit is adapted to perform the steps of the method of the present technology.

[0050] - Optionally, a communication module, for example a WiFi module, is configured to transmit the processed echo signals to a remote computing unit or display unit. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The present invention will be better understood and objects other than those set forth above will become apparent when considering the following detailed description thereof. This description refers to the accompanying drawings, in which:

[0052] FIG1 shows a schematic spectrum with spectral components of a subsequently transmitted probe signal according to a method of the prior art;

[0053] Figure 2 shows a schematic spectrum of spectral components of a subsequently transmitted detection signal in a method according to an embodiment of the present invention;

[0054] Figure 3 A circuit block diagram of an apparatus for detecting underground structures according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0055] Examples of prior art

[0056] In a conventional SFCW GPR, multiple probe signals are sent sequentially into the underground structure, one after the other. Each probe signal typically consists of only one spectral component, which means that each probe signal has a single frequency, for example, a single sinusoidal signal with a certain duration. This is depicted in FIG1 , which shows a schematic overall spectrum of all probe signals: In step 1, a first low-frequency probe signal with, for example, 30 MHz is sent into the underground structure. Afterwards, in step 2, a second probe signal is sent at a slightly higher frequency (for example, 34 MHz). These steps of sending probe signals with increasing frequencies are repeated until the last step, in which a high-frequency probe signal with, for example, 4030 MHz is sent. All probe signals together form a sweep across the entire bandwidth, for example between 30 and 4030 MHz in this example. The use of low and high frequencies improves the resulting image of the underground structure in terms of penetration depth (i.e., the depth to which the structure can be probed downward) and resolution (i.e., the size of the smallest feature that can be resolved), respectively.

[0057] For proper data acquisition and imaging, the entire sweep needs to be sent into the subsurface structure and its echoes received when the antenna has not moved more than the resolution. Obviously, this condition sets an upper limit on the acquisition speed (i.e., the speed at which the antenna moves relative to the subsurface structure).

[0058] In the example above, 1000 steps are required to cover the total bandwidth of the sweep of 4000MHz, with a frequency step size of 4MHz. Assuming that the duration of each probe signal (also called dwell time) is 2μs, and the interruption duration (also called lock time) is 5μs, the total time required for the entire sweep is 7000μs. For a desired resolution of, for example, 20mm, the antenna may not be moved faster than about 10km / h, especially at walking speed.

[0059] On the other hand, acquisition can nowadays be conveniently performed by mounting the antenna on a car or a drone. In that case, an acquisition speed of about 100 km / h may be expected. This speed is 10 times higher than that achievable with prior art methods. In other words, the total time of one sweep will need to be reduced by a factor of 10, while maintaining other parameters (in particular resolution, penetration depth, dwell time and lock time) unchanged.

[0060] Example Method

[0061] The problem of limited acquisition speed is solved by the method according to an embodiment of the present invention, such as Figure 2As shown in the schematic spectrum of . Again, one probe signal is sent at each step, sequentially from step 1 to the last step. However, now each probe signal includes multiple (for example, at least 10) spectral components. The frequency step size between adjacent spectral components is again fixed, for example fixed to 4 MHz, so as to maintain the same penetration depth and resolution as above. By sending 10 spectral components at a time (i.e., within the same probe signal), the total time of the sweep is reduced by a factor of 10. Therefore, the maximum acquisition speed for a given resolution is increased by a factor of 10. In the example above, the maximum acquisition speed is now approximately 100 km / h, which is very suitable for surveying large areas, for example, by car or drone.

[0062] In general, sending multiple spectral components (also called subcarrier signals) concurrently in each probe signal increases the maximum acquisition speed by a factor equal to the number of spectral components. In order for such concurrent sending of subcarrier signals to ultimately produce images or data of a structure comparable to conventional SFCW (e.g., in terms of signal-to-noise ratio and low intermodulation), the subcarrier signals meet certain conditions. These conditions are described in the "Orthogonal Spectral Components" and "Non-Zero Initial Phase Shift" sections above.

[0063] Example Device

[0064] Figure 3 A circuit block diagram of an example device for detecting underground structures is shown. The device is configured to perform a method according to an embodiment of the present invention. Therefore, in the following, aspects related to the device and the method are described.

[0065] The device comprises a digital processing unit 1 and an analog processing unit 2. In the digital processing unit 1, a plurality of modulated signals, for example ten modulated signals, are generated and processed by means of digital components 11 to 15 described in detail below. In a digital-to-analog converter (DAC) 20 with a clock source 21, the digital modulated signals are converted to the analog domain and then further processed by analog components 22 to 28. In particular, a detection signal comprising a plurality of (for example, 10) subcarrier signals is generated by mixing (or in other words frequency converting) the modulated signals with different carrier signals from analog oscillators 23 and 26. The detection signal is then sent into the underground structure by means of a transmitting antenna 29. Components 11 to 28 can be classified as a detection signal generator.

[0066] The echo signal is received from the underground structure by means of a receiving antenna 30. In general, for example in a monostatic radar, the receiving antenna 30 and the transmitting antenna 29 can coincide. The echo signal is processed with analog components 32 to 36. In particular, the echo signal is mixed again with different carrier signals from the analog oscillators 23 and 26, or in other words frequency converted. The frequency-converted echo signal is converted to the digital domain by an analog-to-digital converter (ADC) 37. The echo signal is then further processed by analog components 40 to 47. Components 32 to 47 can be classified as an echo signal processor.

[0067] The device also includes a control unit 48 configured to control the detection signal generator and the echo signal processor. Advantageously, the device also includes a communication unit 49, such as a WiFi module. The communication unit 49 is configured to transmit the processed echo signal to a remote computing device or a display device. The transmission of the processed echo signal can be performed with the output signal of any one of the components 43 to 47 (particularly the component 44), which means that further processing steps are performed by the remote computing device. In addition, the communication unit 49 can also be configured to receive control commands from the remote computing device, such as for starting or ending data acquisition, or for setting parameter values ​​(such as the lower limit, upper limit or frequency step within the entire frequency range of the detection signal subsequently sent, and other parameters that will become clear from the following description).

[0068] Probe signal generation

[0069] exist Figure 3 In the example device of , the modulation signal is generated by a plurality of numerically controlled oscillators (NCOs) in an oscillator unit 11. Alternatively, the NCO may be replaced by an inverse discrete Fourier transform (IDFT), which would replace components 11, 12 and 13.

[0070] As input, the NCO receives the frequency ω n , frequency offset θ n and the initial phase shift α n , where n = 1, 2, ..., N, where N is the number of modulation signals and thus the number of subcarrier signals in each probe signal, for example 10. Modulation frequency ω n The frequency step between adjacent modulation frequencies may be 4 MHz in the range between 40 and 80 MHz. The oscillator unit 11 typically generates a frequency in the form of exp(i((ω n +θ n )t)+α n ) of a digital modulation signal, i.e., a sine and / or cosine signal. Alternatively, if the modulation signal does not change during data acquisition, the discrete value or the sum of the discrete values ​​of the modulation signal can be pre-calculated and stored in a table in the internal memory of the device.

[0071] Advantageously, the initial phase shift α n is non-zero and is different for at least some (particularly all) different n. This has the advantage that the ratio of the maximum detection signal amplitude to the average detection signal amplitude can be kept low over time. Thus, the transmit power of the device can be kept low while maintaining a high signal-to-noise ratio and low intermodulation.

[0072] It has been found that the initial phase shift α n A feasible choice can be to generate α randomly in the interval 0 to 2π n To obtain. For a given N, a random number α can be generated n , and the above low maximum amplitude criterion can be checked in advance. Then, α n The values ​​of may be stored as a table in the device's internal memory and used to generate the detection signal.

[0073] Similarly, the modulation frequency ω n The frequency offset can be shifted by θ n In the case of a constant frequency step size Δω between adjacent modulation frequencies, the different θ n Applied to different frequenciesω n The total cycle time of the detection signal, that is, the total periodicity, can be reduced. However, the frequency offset θ n It should not be too large, for example bounded by the interval (-0.5, +0.5) × Δω, so as not to overlap with adjacent modulation frequencies. In particular, the frequency offset can be 400kHz in order to maintain the orthogonality criterion defined previously. In particular, all frequency offsets θ n can be the same, θ n =θ.

[0074] Advantageously, the frequency offset θ n is the same for all modulation signals in a detection signal. Further advantageously, the frequency offset θ n The detection signals for all the detections in one cycle are the same. n Can be applied to different cycles to reduce overall cycle time.

[0075] Generally speaking, it is necessary to at least temporarily store or know the frequency ω involved in the modulation n , frequency offset θ n and the initial phase shift α n The detection signal generates an input value for subsequent processing of the echo signal, in particular in the digital mixer 41.

[0076] In the next step, the switching matrix 12 is applied to the digital modulation signal. Applying the switching matrix 12 can turn off the modulation signal of a certain modulation frequency. This can be useful for reducing the number of subcarrier signals in certain detection signals of the sweep, for example, from N=10 to 5 or only 1 subcarrier signal per detection signal, while maintaining the number of subcarrier signals in other detection signals of the same sweep at N. In this way, the peak power can be reduced. Moreover, less processing power is required and ultimately less data needs to be transmitted, for example, through the communication unit 49 (for example, a WiFi unit after the serializer 42). For example, it can be useful to use a frequency step size Δω=4MHz for low-frequency detection signals (for example, between 40 and 1500MHz) and a frequency step size Δω=8MHz for higher frequency detection signals (for example, above 1500MHz). In this case, when providing modulation signals for detection signals above 1500MHz, the switching matrix 12 is zeroed every two modulation signals.

[0077] Then, a summation of all digital modulated signals is performed by a frequency band combiner 13. Subsequently, a dynamic amplitude correction 14 may be performed on the combined signal as already described above.

[0078] In addition, still in the digital domain, a gating 15 is applied to the combined signal. Gating advantageously applies a smooth time window function to the combined signal. For example, the time window can be a Gaussian function, and it reduces artifacts typically caused by applying a hard window (such as a sudden opening or closing of the signal).

[0079] The combined modulated signal is now converted to the analog domain by the DAC 20 which receives a clock signal having a clock frequency from the clock oscillator 21. Figure 3 In the example device of , the clock frequency is 200 MHz, resulting in an analog signal which is then bandpass filtered by filter 22 with a frequency range of 120 to 160 MHz.

[0080] This signal is then mixed with a first carrier signal of a first carrier frequency generated by an analog oscillator 23. The mixing, in this case up-conversion, is performed by means of an analog mixer 24. This mixing process is also referred to as "heterodyning", wherein in particular only the "upper heterodynes" (i.e. at a frequency equal to the sum of the first carrier frequency and the modulation frequency) are used, while the "lower heterodynes" (i.e. at a frequency equal to the difference between the first carrier frequency and the modulation frequency) are discarded (e.g. filtered out). In general, mixing a signal towards a higher frequency is referred to as "up-conversion", while mixing a signal towards a lower frequency is referred to as "down-conversion".

[0081] In the example device, the first carrier frequency is 3600 MHz, resulting in a first mixed signal which is then bandpass filtered by the filter 25 in the frequency range of 3720 to 3760 MHz. The first mixed signal is then down-converted by the analog mixer 27. To this end, the analog mixer 27 receives a second carrier signal from the analog oscillator 26. The analog oscillator 26 is a variable frequency oscillator configured to generate a second carrier signal having a second carrier frequency, wherein the second carrier frequency is between 3800 and 7200 MHz. The oscillator 26 is advantageously controlled by the control unit 48. In particular, the second carrier frequency is changed in steps of 40 MHz (i.e., the bandwidth of one probe signal) between the subsequent detection signals from 3800 to 7200 MHz in general. This results in a second mixed signal. The second mixed signal in particular has a bandwidth of 40 MHz and includes a subcarrier frequency of Δω=4 MHz. The swept subsequent second mixed signal is low-pass filtered by the filter 28. The filter 28 advantageously has a cut-off frequency of 3440 MHz. The filtered second mixed signal which constitutes the actual detection signal accordingly has a frequency in the frequency range between 40 and 3440 MHz. The detection signal is then transmitted by the transmitting antenna 29 .

[0082] As an alternative to the exemplary device, the mixing of the analog modulation signal to the detection signal, i.e. the modulation, can be performed with the aid of only one analog mixer. Such up-conversion in one step requires more filtering due to the need to remove harmonics from the signal. In contrast, in Figure 3 In the example device of , “image frequencies” (eg, up-heterodyne) are automatically discarded by sequentially applying several mixing steps.

[0083] Echo signal processing

[0084] The echo from the underground structure is received as an analog echo signal by the receiving antenna 30 and filtered by a low-pass filter 32, in particular in the frequency range of the detection signal, in particular with a cut-off frequency of 3440 MHz.

[0085] Further processing of the echo signal in the analog processing unit 2 generates an image from the detection signal as described above. In general, the detection signal is down-converted to the frequency range of the modulation signal, in particular to the range of 40 to 80 MHz. This can be done in one step, or as Figure 3 This is done in a few steps, as in the example device of FIG.

[0086] In the exemplary device, the echo signal is mixed, in particular frequency converted, with the second carrier signal from the second analog oscillator 26 by means of a mixer 33. This means that the echo signal is up-converted to a first mixed echo signal, which is then filtered by a bandpass filter 34 in a frequency range between 3720 and 3760 MHz. The first mixed echo signal is then down-converted by mixing with the first carrier signal from the first oscillator 23 by means of a mixer 35. The resulting second mixed echo signal is filtered by a bandpass filter 36 in a frequency range between 120 and 160 MHz. The second mixed echo signal is then converted to the digital domain by an ADC 37, which receives a clock signal having a clock frequency from the clock oscillator 21.

[0087] In the digital processing unit 1, a temporal weighting 40 is applied to the echo signal. In particular, this is used to reduce spectral leakage of a subsequent discrete Fourier transform (DFT). The echo signal is then input to a digital mixer 41, where it is advantageously transformed into the frequency domain, for example by applying a DFT. The different spectral components of the echo signal corresponding to the frequency of the modulation signal (e.g., 10 if N=10) are separated and compared with respect to an initial phase shift α. n Make corrections.

[0088] Additionally, the offset or average value in each spectral component can be removed. The average value after the mixing stage reduces the signal to one complex coefficient per frequency band. It also defines the equivalent noise bandwidth (ENBW) of the system, which is given by the duration of the received signal and the shape of the gate 15 and the weighting 40.

[0089] In order to obtain a processed echo signal comparable to that of a conventional (i.e. non-multiplexed) SFCW method, the different spectral components need to be "serialized", i.e. arranged in time order. This is done by serialization 42, for example by applying a delay matrix comprising a time shift Δt for each of the N spectral components n As a result of the serialization 42, the echo signal has the form of a sweep of consecutive signals, wherein each subsequent signal consists of only one spectral component and differs from the frequency of the previous signal by a frequency step Δω.

[0090] Further processing of the echo signal advantageously includes correcting the echo signal for the device response, i.e., correction for changes in amplitude and / or phase due to components of the device. To this end, a calibration signal 31 may be measured between the transmitting antenna 29 and the receiving antenna 30, e.g., before each sweep of the detection signal. The calibration signal 31 may be used to correct the echo signal in the calibration 43. In particular, drifts of the signal or device response over time may be corrected by using the current calibration signal 31 and historical calibration signals, which may be recorded, for example, by the manufacturer of the device.

[0091] Additionally, spatial averaging 44 may be applied to the echo signals. Spatial averaging increases the signal-to-noise ratio, which reduces the amount of data. Since, for example, the communication unit 49 will store or transmit the echo signals, spatial averaging 44 is advantageously applied before the transmission of the data.

[0092] It is common practice to apply further signal processing steps to the echo signal, such as at least one of weighting 45, zero padding 46 and inverse Fourier transform 47. These steps are advantageously performed on a remote computing device. As a result of the inverse Fourier transform 47, the echo signal is converted into a time domain waveform, also referred to as an A-scan in the context of GPR.

[0093] In general, it is advantageous to use dedicated hardware not only for the analog processing unit 2 but also for the digital processing unit 1. The digital processing unit 1 is advantageously one or more field programmable gate arrays (FPGAs) configured to perform one or more of the described processing steps. The use of FPGAs facilitates efficient detection signal generation and / or echo signal processing, in particular real-time processing.

[0094] While there are shown and described presently preferred embodiments of the invention, it is to be distinctly understood that the invention is not limited thereto but may be otherwise variously embodied and practiced within the scope of the appended claims.

Claims

1. A method for detecting underground structures, comprising the following steps: sending electromagnetic waves into the structure, receiving an echo of the electromagnetic wave from the structure, processing the echoes to derive the internal features of the structure, wherein the step of sending electromagnetic waves into the structure comprises subsequently sending a plurality of electromagnetic detection signals having different frequency spectra into the structure, wherein each detection signal comprises at least two non-zero spectral components, The step of receiving the echo includes receiving an echo signal for each detection signal, wherein the step of processing the echoes comprises determining at least one amplitude and phase for each echo signal, wherein at least two of the spectral components have a non-zero initial phase shift, The initial phase shifts of the at least two spectral components make the maximum amplitude of the detection signal smaller than the maximum amplitude of an assumed detection signal having the same spectral components but with the initial phase shift of the spectral components being zero.

2. The method according to claim 1, The spectral components are concurrent subcarrier signals with different subcarrier frequencies.

3. The method according to claim 1, The spectral components are located in non-overlapping frequency bands.

4. The method according to claim 1, Wherein the at least two spectral components are orthogonal to each other.

5. The method according to claim 4, The frequencies of adjacent subcarriers differ from each other by a subcarrier spacing.

6. The method according to claim 4, The subcarrier spacing satisfies the orthogonality condition.

7. The method according to claim 4, The subcarrier spacing is between 1 and 20 MHz.

8. The method according to claim 7, The subcarrier spacing is 4MHz.

9. The method according to claim 1, The frequency spectrum of the detection signal lies between 10 and 8000 MHz.

10. The method according to claim 9, The frequency spectrum of the detection signal lies between 40 and 3440 MHz.

11. The method according to claim 1, The total frequency spectrum of the plurality of electromagnetic detection signals subsequently transmitted has a width of at least 500 MHz.

12. The method according to claim 11, The total frequency spectrum of the plurality of electromagnetic detection signals subsequently transmitted has a width of at least 1000 MHz.

13. The method according to claim 11, The total frequency spectrum of the plurality of electromagnetic detection signals subsequently transmitted has a width of at least 2000 MHz.

14. The method of claim 1, The maximum amplitude of the detection signal is at least 10%, at least 25% or at least 50% smaller than the maximum amplitude of the assumed detection signal.

15. The method of claim 1, Each of the detection signals comprises between 3 and 20 non-zero spectral components.

16. The method of claim 15, Each of the detection signals comprises between 5 and 15 non-zero spectral components.

17. The method of claim 15, Each detection signal includes 10 non-zero spectral components.

18. The method of claim 1, Additionally, the method includes the step of generating a detection signal by frequency converting at least two modulation signals of different modulation frequencies with at least one carrier signal.

19. The method of claim 18, Wherein the step of generating the probing signal comprises changing a carrier frequency of the carrier signal between subsequent probing signals.

20. The method of claim 18, The modulation frequency is the same for all detection signals.

21. The method of claim 18, The step of generating a detection signal includes generating at least two digital modulation signals.

22. The method of claim 21, The step of generating a detection signal includes generating at least one analog carrier signal.

23. The method of claim 21, The at least two digitally modulated signals have a non-zero initial phase shift.

24. The method of claim 18, The modulation frequency is lower than 100MHz.

25. The method of claim 24, The modulation frequency is in the range between 40 and 80 MHz.

26. The method of claim 24, The modulation frequency varies based on the subcarrier spacing.

27. The method of claim 18, The carrier frequency of the carrier signal is at least 100 MHz.

28. The method of claim 21, The step of generating the detection signal includes obtaining at least two analog modulation signals by performing digital-to-analog conversion on the at least two digital modulation signals, and mixing the at least two analog modulation signals with at least one mixing signal from at least one analog oscillator by means of at least one analog mixer (24, 27).

29. The method of claim 28, The step of generating the detection signal comprises sequentially mixing the analog modulation signal with mixing signals from a plurality of analog oscillators by means of a plurality of analog mixers (24, 27).

30. The method of claim 28, The step of generating the detection signal comprises applying up-conversion to the analog modulation signal.

31. The method of claim 30, The step of generating the detection signal comprises applying an up-conversion and a subsequent down-conversion to the analog modulated signal.

32. The method of claim 21, The step of generating a digital modulation signal includes turning off a predetermined modulation signal.

33. The method of claim 1, The first detection signal in the detection signal includes a different number of non-zero spectral components than the second detection signal in the detection signal.

34. The method of claim 1, The total duration of each detection signal does not exceed 2μs.

35. The method of claim 1, The envelope of each detection signal has a rise time and a fall time of non-zero duration.

36. The method of claim 35, Therein the envelope is shaped as a Gaussian function.

37. The method of claim 1, The step of transmitting electromagnetic waves into the structure is performed with the aid of an antenna.

38. The method of claim 37, The step of transmitting electromagnetic waves into the structure is performed by a higher frequency sub-antenna and a lower frequency sub-antenna.

39. The method of claim 28, The step of processing the echo signal comprises obtaining a mixed echo signal by mixing the echo signal with the at least one mixing signal from the at least one analog oscillator by means of at least one further analog mixer (33, 35).

40. The method of claim 39, The step of processing the echo signal comprises sequentially mixing the echo signal with mixing signals from a plurality of analog oscillators by means of a plurality of further analog mixers (33, 35).

41. The method of claim 39, The step of processing the echo signal comprises applying a down conversion to the echo signal.

42. The method of claim 41, The step of processing the echo signal comprises applying an up-conversion and a subsequent down-conversion to the echo signal.

43. The method of claim 39, The step of processing the echo signal includes converting the mixed echo signal into a digital echo signal.

44. The method of any one of claims 1 and 43, The step of processing the echo signal comprises correcting the digital echo signal for an initial phase shift of the at least two spectral components.

45. The method of claim 1, The step of processing the echo signal includes obtaining a serialized echo signal by serializing the at least two non-zero spectral components.

46. ​​The method of claim 45, The step of processing the echo signal comprises obtaining a serialized echo signal by serializing the at least two non-zero spectral components using the digital echo signal.

47. The method of claim 1, The step of processing the echo includes deriving at least one characteristic of the underground structure from an amplitude difference and / or a phase difference between the echo signal and the detection signal.

48. An apparatus for detecting underground structures, comprising: antenna, a detection signal generator, for generating an electromagnetic detection signal to be sent into the underground structure via an antenna, an echo signal processor for processing the echo signal received from the underground structure via the antenna, a control unit (48) for controlling the detection signal generator and the echo signal processor, The device is configured to perform the method as claimed in claim 1.

49. The apparatus of claim 48, comprising: a first antenna configured to transmit a detection signal into the underground structure, and The second antenna is configured to receive the echo signal received from the underground structure.

50. The apparatus of claim 49, The first antenna and the second antenna both include a lower frequency sub-antenna and a higher frequency sub-antenna.

51. The apparatus of claim 48, The detection signal generator includes: - a digital signal generator configured to generate at least two digital modulation signals, - a digital-to-analog converter (20) configured to convert the digital modulation signal into an analog modulation signal, - at least one analog signal generator configured to generate at least one carrier signal, and - at least one analog mixer (24, 27) configured to generate a detection signal by mixing the analog modulation signal with the at least one carrier signal.

52. The apparatus of claim 51, The echo signal processor includes: at least one further analog mixer (33, 35) configured to generate a mixed echo signal by mixing the echo signal with the at least one carrier signal from the at least one analog signal generator, - an analog-to-digital converter (37) configured to convert the mixed echo signal into a digital echo signal, and - a digital processor configured to derive the at least two non-zero spectral components from the digital echo signal.

Images