An improved metal detector
A two-stage signal processing method for metal detectors enhances target detection accuracy and reduces false alarms by maintaining sensitivity to targets like gold nuggets and distinguishing them from soil components, using pre-whitening and filtering techniques.
Patent Information
- Application Number
- PCT/AU2026/050478
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-08-12
- Filing Date
- 2026-05-13
- Publication Date
- 2026-11-19
AI Technical Summary
Existing metal detectors struggle to maintain high sensitivity to targets of interest, such as gold nuggets, while reducing false alarms from magnetic soils and inhomogeneous soil components like 'hot rocks', especially in gold-rich regions with varying magnetic properties.
A two-stage signal processing method that generates a first signal sensitive to targets of interest and insensitive to soil components, and a second signal that discriminates between targets and soil components, using pre-whitening and filtering techniques to produce distinct indicator outputs.
Enhances target detection accuracy and reduces false alarms by maintaining high sensitivity to targets of interest and distinguishing them from soil components, improving operational efficiency and productivity in mineralized soils.
Smart Images

Figure AU2026050478_19112026_PF_FP_ABST
Abstract
Description
AN IMPROVED METAL DETECTORTECHNICAL FIELD
[0001] The present disclosure relates to a metal detector.BACKGROUND
[0002] The general forms of most metal detectors are either hand-held battery-operated units, conveyor-mounted units, or vehicle-mounted units. Examples of hand-held products include detectors used to locate gold, explosive land mines or ordnance, coins and treasure. Examples of conveyormounted units include fine gold detectors in ore mining operations, and an example of a vehiclemounted unit includes a unit to locate buried land mines.
[0003] These metal detectors usually, but not necessarily, comprise transmit electronics generating a repeating transmit signal cycle of a fundamental period, which is applied to an inductor, for example a transmit coil, which transmits a resulting varying magnetic field, sometimes referred to as a transmit magnetic field.
[0004] These metal detectors may also comprise receive electronics that processes a receive signal from a measured receive magnetic field, during one or more receive periods during the repeating transmit signal cycle, to produce an indicator output signal, the indicator output signal at least indicating the presence of at least a metal target within the influence of the transmit magnetic field.
[0005] During the processing of the receive signal, the receive signal is either sampled, or demodulated, to produce one or more target channels, the one or more target channels being further processed to produce the indicator output signal.
[0006] Metal detectors may be classified into two groups depending on their transmit signal and demodulation (or sampling) functions, namely frequency-domain metal detectors or time-domain metal detectors. Frequency-domain metal detectors may be thought of as having signal processing that includes demodulators measuring at specific frequencies of their transmit signal whilst time-domain metal detectors may be thought of as having demodulators measuring at specific periods of time usually following transitions in a transmit signal. Examples of time-domain metal detectors are pulseinduction (PI) metal detectors, or Pl-like metal detectors with periods of alternating constant transmitted magnetic fields with rapid magnetic field transitions connecting the alternate periods of constant transmitted magnetic fields. Examples of frequency-domain metal detectors include one ofthe most common forms of metal detectors, the single frequency sine wave transmitting type and, multi-frequency metal detectors that usually transmit and receive several different frequencies.
[0007] Magnetic soils in gold-rich regions are often characterised by a high concentration of magnetic minerals, such as magnetite, that may be naturally enriched by the same hydrothermal and weathering processes which concentrate gold. These soils typically exhibit both an instantaneous magnetic response and a viscous remanent magnetisation (VRM) response. Under the alternating magnetic field of a metal detector, reversible domain-wall bowing in multi-domain particles and Neel rotations in single-domain particles produce a frequency-dependent induced magnetisation that appears almost instantaneously compared to the period of a typical metal detector cycle. Superimposed on this is a slow, time-dependent response from VRM. VRM arises as magnetic mineral grains gradually align their magnetic moments with an applied magnetic field. This process is thermally activated and depends on factors like grain size and temperature. Magnetic soils often contain a range of magnetic particle sizes which exhibit a broad range of thermal relaxation times but mostly exhibit an approximately log-uniform distribution of relaxation times. The prevalence of this approximately loguniform distribution of relaxation times leads to a highly predictable, characteristic soil response to the applied magnetic field from a metal detector. Metal detectors may therefore utilise signal processing techniques relying on this characteristic soil response to significantly reduce the response from the soil relative to the response from targets of interest. Small variations in the approximately log-uniform distribution of relaxation times are encountered in practice. The variation is well-described by the soil having a small, variable amount of a higher-order component that results from an approximately log-linear distribution of relaxation times. Thus, the soil is more accurately described as a linear combination of the log-uniform and log-linear distributions of relaxation times, for example, as described in US5576624. Variation in the ratio of the log-linear component compared to the loguniform component is observed at both small (<10cm) and large (>lm) spatial scales as a metal detector traverses a soil. Additionally, many isolated rocks may be found which have log-linear to loguniform ratios that are significantly different than the soil they are located in. These rocks are commonly referred to as “hot rocks” in the metal detecting community, being readily detectable by metal detectors that are calibrated to reduce the response of the bulk surrounding soil matrix rather than that of the hot rock.
[0008] Known techniques exist for significantly reducing the response from the VRM soil relative to the response from targets of interest. US9366779 describes a technique for utilising a model of the VRM soil response to essentially create projections of the receive signal that are orthogonal to the modelled VRM soil response. US5576624 describes a technique for a PI metal detector wherein responses from two different time periods are combined to reduce the response from the soil.US8106770 further describes a technique wherein at least three different frequency components are combined to reduce a primary response from the soil in the form of log-uniform resistive signalcomponents and a secondary response from the soil in the form of log-linear frequency-dependent resistive signal components. In practice, the technique described in US8106770 significantly reduces the false alarm rate from isolated rocks that may have substantially different magnetic properties compared to the surrounding soil matrix, and it may also significantly reduce the false alarm rate when the operator moves between different areas of the soil that may have substantially different magnetic properties compared to each other. However, the technique of US8106770 significantly reduces the sensitivity of a metal detector to gold nuggets, and particularly those gold nuggets which are either very small or very large in size.
[0009] Known techniques exist for classifying between different classes of targets of interest and undesired objects. For example, US5506506 describes a method for a single-frequency metal detector that utilises the phase of the metal detector response to classify between different classes of targets of interest and undesired objects, for example, discriminating between ferrous and non-ferrous targets. AU2024902579 describes a method for a multi-frequency metal detector that utilises comparisons between multiple channels, including both resistive and reactive channels, to classify between different classes of targets of interest and undesired objects. Neither of these patents describes the combination of a detection output indicator and a classification output indicator that are both insensitive to at least the log-uniform VRM soil components, and with the classification output indicator discriminating between targets of interest and soil components.
[0010] US9429674 describes a method whereby an indicator output is produced which has the effects of signals from the soil reduced therein or removed therefrom, but does not teach the combination of producing a high-sensitivity detection output indicator with some but not all of the effects of signals from the soil reduced therein or removed therefrom to maintain high sensitivity to targets of interest, or producing a classification indicator output that discriminates between targets of interest and other different soil components.
[0011] The present disclosure offers an alternative over existing prior art.SUMMARY OF THE DISCLOSURE
[0012] According to a first aspect of the present disclosure, there is provided a method for detecting one or more targets of interest in a soil with a metal detector, the method comprising: generating a transmit magnetic field for transmission into the soil based on a transmit signal; receiving a receive magnetic field to induce a receive signal; processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and insensitive to a first component of the receive signal due to the soil; processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targetsof interest and an unwanted second component different from the first component; and producing at least one indicator output signal based on the first signal and the second signal.
[0013] In one form, the unwanted second component is due to the soil, ferrous objects, or both.
[0014] In one form, the second signal is less sensitive to the one or more targets of interest than the first signal.
[0015] In one form, the method further comprises: producing a first indicator output signal based on the first signal; and producing a second indicator output signal based on the second signal.
[0016] In one form, the first indicator output signal and the second indicator output signal are presented to a user of the metal detector simultaneously.
[0017] In one form, the first component comprises an instantaneous reactive magnetic response, or a log-uniform distributed magnetic soil VRM component, or both.
[0018] In one form, the second signal discriminates between the one or more targets of interest and a second component due to the soil, the second component being different from the first component.
[0019] In one form, the second signal discriminates between the one or more targets of interest that are non-ferrous and a second component due to ferrous objects.
[0020] In one form, the second component comprises a log-linear distributed magnetic soil VRM component, an instantaneous reactive magnetic response, or both.
[0021] In one form, the second component comprises a resistive component that increases approximately linearly with frequency.
[0022] In one form, the step of processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and is insensitive to a first component of the receive signal due to the soil comprises one or both of the following: producing a projection of the receive signal that is orthogonal to the first component due to the soil in a time domain or frequency domain; and prewhitening based on a historical noise measurement.
[0023] In one form, the step of processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and a second component comprises: pre-whitening based on a historical noise measurement.
[0024] In one form, the step of producing at least one indicator output signal based on the first signal and the second signal comprises: comparing a likelihood of a first target of interest and noise being present with a likelihood of noise only being present; and comparing a likelihood of the first target of interest being present with a likelihood of the second component being present.
[0025] In one form, the first indicator output and the second indicator output are in the form of one or more of visual, audio, or haptic indicators.
[0026] In one form, the step of producing at least one indicator output signal based on the first signal and the second signal comprises applying a high-pass or band-pass filter with at least 6dB of attenuation at 0 Hz.
[0027] In one form, the first indicator output and the second indicator output are simultaneously presented via a user interface to a user when the metal detector is used in a non-motion mode.
[0028] In one form, the one or more targets of interest comprise gold nuggets weighing between 0.01g and 100g each.
[0029] In one form, the step of processing the receive signal to produce the second signal comprises combining low, medium, and high frequency components of the receive signal to produce two channels that are both insensitive to the first component; and wherein the first of the two channels is insensitive to a particular target of interest and the second of the two channels is sensitive to the particular target of interest.
[0030] According to another aspect of the present disclosure, the metal detector is a handheld metal detector.
[0031] According to another aspect of the present disclosure, there is provided a metal detector configured to perform the method of the first aspect.BRIEF DESCRIPTION OF DRAWINGS
[0032] Embodiments of the present disclosure will be discussed with reference to the accompanying drawings wherein:
[0033] Figure 1 depicts a general form of the present disclosure;
[0034] Figure 2 depicts an embodiment of the present disclosure;
[0035] Figure 3 depicts another embodiment of the present disclosure; and
[0036] Figures 4A to 4C depict an exemplary implementation of the present disclosure.DESCRIPTION OF EMBODIMENTS
[0037] In situations where soil components associated with hot rocks or patches of magnetically inhomogeneous soil are present but relatively rare, it is beneficial to process the receive signal so as to create indicator outputs that are insensitive to the soil components present in the bulk surrounding soil matrix without concern for reducing the false alarm rate from the soil components that are associated specifically with the hot rocks or magnetically inhomogeneous soil in order to maintain high sensitivity to targets of interest. Indicator outputs that are insensitive to the additional soil components are generally much less sensitive to targets of interest as the additional soil components have some similarity to most targets of interest. Reducing the sensitivity to one reduces the sensitivity to the other. To reduce the false alarm rate, a second-stage classifier is used that discriminates between targets of interest and responses due to the higher-order soil components. It is noted that the signal-to-noise ratio must be sufficiently high for the second-stage classifier to be reliable, and typically higher than necessary for the first-stage detector to declare that an anomaly is present. However, it is common to find hot rocks that produce a sufficiently large response for the second-stage classifier to be useful in identifying these hot rocks before time and physical effort is expended to retrieve them from the surrounding soil. Similarly, the second-stage classifier declaring that a target of interest is present provides the user with the increased confidence to persist during a time-consuming and laborious target retrieval process.
[0038] It is of further benefit in speeding up the operation of the metal detector that the second-stage classification result be available and presented to the user instantaneously during normal operation of the metal detector rather than requiring the user to activate a special mode or perform a special interrogation process. Speeding up the operation of the metal detector may increase the productivity of the metal detector by increasing the rate of area of ground searched for gold.
[0039] It is of further benefit in increasing the accuracy of the classification result that at least a high-pass filter or band-pass filter is used to attenuate the slowly time-varying response from the bulk surrounding soil matrix, such as the filters that are applied in metal detector modes commonly referred to as motion modes (as opposed to non-motion modes or pinpointing modes). The slowly time-varying response from the bulk soil is a source of noise that is typically present when a potential buried target of interest is encountered. Attenuating this slowly time-varying response may thus improve the signal-to-noise ratio.
[0040] Figure 1 depicts a general form of the present disclosure for detecting one or more targets of interest in a soil with a metal detector. The first step 1 comprises the step of generating a transmit magnetic field for transmission into the soil based on a transmit signal. The next step 3 comprises the step of receiving a receive magnetic field to induce a receive signal. These two steps may take many forms known to a person skilled in the art, as long as the transmitter transmits a magnetic field to excite a target based on a transmit signal produced by transmit electronics of the metal detector and the receiver receives signals indicative of the presence of the target within the influence of the transmitted magnetic field. The transmitter and receiver may take many different forms depending on the designs and the implementations of the metal detector. The transmit signal may take many forms known to a person skilled in the art as long as it is suitable for the operation of the metal detector. The one or more targets of interest may be gold nuggets, weighing between 0.01g and 100g each. The one or more targets of interest may also be explosive land mines or ordnance, coins and treasure.
[0041] The next step 5 comprises the step of processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and insensitive to a first component of the receive signal due to the soil. The term “sensitive” would be understood by a person skilled in the art to mean able to detect. In the context of “a first signal that is sensitive to the one or more targets of interest”, it simply means the first signal contains or carries information in relation to the one or more targets of interest. If the first signal is linked to an earphone, for example, a user of the metal detector would be able to hear a “beep”, indicating that there is a high likelihood that the metal detector has detected a presence of the one or more targets of interest. The term “insensitive” on the other hand means that the first signal contains or carries relatively low information of the first component of the receive signal due to the soil. Alternatively, “insensitive” in this context may be understood to mean that the overall noise in the first signal increases by no more than a factor of two due to the response from the first component of the receive signal due to the soil.
[0042] For example, the first component comprises an instantaneous reactive magnetic response of the soil or instantaneous magnetic response from a ferrous object.
[0043] For another example, the first component comprises both an instantaneous reactive magnetic response of the soil and a log-uniform distributed magnetic soil VRM component. In other words, the first signal is not affected adversely by the presence of unwanted signals due to the soil. In one embodiment, step 5 of processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and is insensitive to a first component of the receive signal due to the soil comprises: producing a projection of the receive signal that is orthogonal to the first component due to the soil in a time domain or frequency domain.
[0044] In one embodiment, step 5 of processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and is insensitive to a first component of the receive signal due to the soil comprises: pre-whitening based on a historical noise measurement. For example, it may be based on an inverse autocovariance matrix of selected components of the receive signal. An approximation of the inverse autocovariance matrix may alternatively be used in the pre-whitening step.
[0045] The next step 7 comprises processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and an unwanted second component different from the first component. For example, the second component comprises a log-linear distributed magnetic soil VRM component. For another example, the second component comprises an instantaneous reactive magnetic response of the soil. For another example, the second component comprises a resistive component that increases approximately linearly with frequency.
[0046] Figure 2 shows step 7a, as one alternate embodiment of step 7 of Figure 1. Other steps 1, 3, 5, 9 are the same with those of Figure 1. Step 7a comprises processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and a second component due to the soil.
[0047] Figure 3 shows step 7b, as one alternate embodiment of step 7 of Figure 1. Other steps 1, 3, 5, 9 are the same with those of Figure 1. Step 7b comprises processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and a second component due to ferrous objects. This allows discrimination between the one or more targets of interest that are non-ferrous and a second component due to ferrous objects. In one embodiment, the second component comprises an instantaneous reactive magnetic response and a conductive target response; wherein the instantaneous reactive magnetic response is at least an order of magnitude larger than the conductive target response.
[0048] In one embodiment, step 7 of processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and a second component at a higher order than the first component comprises: pre-whitening based on a historical noise measurement, for example, an inverse autocovariance matrix of selected components of the receive signal.
[0049] In one embodiment, step 7 of processing the receive signal to produce the second signal comprises combining low, medium, and high frequency components of the receive signal to produce two channels that are both insensitive to the first component; and wherein the first of the two channelsis insensitive to a particular target of interest and the second of the two channels is sensitive to the particular target of interest.
[0050] There are many unwanted components of the receive signal, such as components due to the soil or ferrous objects. It is tempting to cancel all of them so that the first signal is free of any unwanted signal. It was found that doing so will also remove some of the useful signals representing one or more targets of interest. Thus, the step of processing the receive signal to produce the first signal and second signal is configured such that the components contained within the first signal and second signal are different so that in the absence of disturbance such as hot rock, the first signal is more sensitive to the intended targets, while the second signal is useful to determine the likelihood of the first signal being affected by disturbance such as hot rock or ferrous objects.
[0051] In one embodiment, the second signal is less sensitive to the one or more targets of interest than the first signal.
[0052] The next step 9 comprises producing at least one indicator output signal based on the first signal and the second signal. The first indicator output and the second indicator output may take the form of one or more of visual, audio, or haptic indicators etc.
[0053] In one embodiment, there are two different indicators. The first indicator output signal is based on the first signal, and the second indicator output signal is based on the second signal. In one embodiment, the first indicator output signal and the second indicator output signal are presented to a user of the metal detector simultaneously. In another embodiment, the first indicator output signal and the second indicator output signal are combined to form a combined output signal presented to a user of the metal detector.
[0054] In one embodiment, the step 9 of producing at least one indicator output signal based on the first signal and the second signal comprises applying a high-pass or band-pass filter with at least 6dB of attenuation at 0 Hz.
[0055] In one embodiment, the step 9 of producing at least one indicator output signal based on the first signal and the second signal comprises: comparing a likelihood of a first target of interest and noise being present with a likelihood of noise only being present; and comparing a likelihood of the first target of interest being present with a likelihood of the second component being present.
[0056] In one embodiment, the first indicator output and the second indicator output are simultaneously presented via a user interface to a user when the metal detector is used in a non-motion mode.
[0057] In one embodiment, the first indicator output is a first audio output, and the second indicator output is a second different audio output that occurs simultaneously with the first audio output. This has the advantage of improving the speed and efficiency of metal detection.
[0058] In one embodiment, the second indicator output is a visual indicator output that is presented on a user interface.
[0059] The disclosure may be applied to any metal detector, including a handheld metal detector, or a conveyor-mounted metal detector, or a vehicle-mounted metal detector.
[0060] To explain the disclosure further, consider the case wherein the user is operating a metal detector in mineralised soil. When an anomaly is found with the metal detector the user wants to decide which among three hypotheses:HoNo target of interest present (noise only).H. A target from target set A is present (the desired target, e.g. a gold nugget).H2A target from target set B is present (rare confusor, e.g. a hot rock).
[0061] Based on a two-stage approach, one may assume that under each hypothesis HL. the data x follows a distribution p((x). The prior probabilities of each hypothesis are P(H0) = n0, P(H ) =P(W2) = 7T2, with 7T0+ 7?! + 7T2= 1.
[0062] During Stage 1, consider a first test statistic, ^(x). optimised for detecting A in noise - that is, for discriminating 7^ from Ho. Formally, we might define:
[0063] In classical Neyman-Pearson detection theory (for a simple hypothesis test ofvs. Wo). the decision rule is:-> " Declare detection of A-like target",-> " No detection".
[0064] Here is a threshold chosen to achieve a desired probability of false alarm. Crucially, Stage 1 does not try to distinguish A from B - it only tries to maximise the probability of detecting Arelative to noise. As a side effect, it may also declare a “detection” if B is present, potentially causing a false alarm from B.
[0065] During Stage 2, the second test statistic, S2(x), is used to discriminate A from B for those x already declared “detection” by Stage 1. One may define:
[0066] If Stage 1 indicates a detection (Sx(x) >?71). then we compare S2(x) to a threshold T]2• For a simple A vs. B test, the Neyman-Pearson (or Bayesian) approach yields:" Declare target A",-> " Declare target B".
[0067] In practice, r / 2may be set to achieve a desired trade-off between incorrectly labelling B as A and incorrectly labelling A as B.
[0068] It is not necessary to explicitly split the decision rule into two stages. Instead, the two stages may be combined into a single-stage decision rule that is mathematically equivalent. However, there are several benefits of a two-stage decision rule.
[0069] Classification typically requires a stronger signal or higher SNR before it is reliable. In a two-stage approach the classification step needs only to be applied after a detection is relatively certain.
[0070] With a human operator in the loop, the operation may monitor the first indicator output driven by the first stage detection test statistic, which tends to be a simple, continuous function of the receive signal and therefore easy to interpret, only gathering more information from the second indicator output driven by the second stage classification test statistic when the first stage detection is certain. Thereby the cognitive load on the operator is reduced.
[0071] The second stage classification calculation may be computationally expensive. Performing this calculation only when the first stage detection indicates a likely target may decrease computational burden.
[0072] For determining likelihood ratios for a multi-frequency metal detector, one may consider a multi-frequency metal detector producing a n-frequency complex signal expressed in terms of its realand imaginary components. We denote the observation (data) vector by x G IR / V 2n. Under each hypothesis:Ho(noise only):x~ (0, C)H (target A present, unknown amplitude aA)X~N aASA, OH2(target B present, unknown amplitude aB)x~N(aBsB, C)
[0073] A pre-whitening step typically involves techniques for cancelling or attenuating soil response components. For example, it may involve taking projections of the data that are orthogonal to expected or measured soil response components in the data or attenuating the data into the directions of the expected or measured soil response components. We note, however, that the soil response components are generally not orthogonal to the expected response from targets of interest, so it is desirable to perform as few such orthogonal projections as possible to maintain high sensitivity to targets of interest.
[0074] For clarity, in this context, pre-whitening involves transforming the data vector x by multiplying by a pre-whitening matrix W. A pre-whitening matrix may be defined asWTW = C-1where C-1is the inverse of the covariance matrix of the data x. In practice, the components of C due to the soil are often so large compared to other noise sources such as electromagnetic interference or electronic noise that the matrix C may be approximated by assuming the components due to the soil are effectively infinite in magnitude, and hence C-1may not be full rank. In essence, applying the prewhitening matrix W may have the effect of creating projections of the data matrix x that are orthogonal to the soil components, thus making the pre-whitened data matrix insensitive to these soil components. In other cases, it may be beneficial to merely attenuate some of the soil components, especially when these soil components are highly correlated with components of the data that have significant discriminating power in terms of the hypothesis tests. Attenuating these soil componentstoo much may have the consequence of incidentally attenuating the components of the data that have significant discriminating power.
[0075] In effect, the pre-whitening process is equivalent to the time-domain methods described in US5576624 and US8106770, changing the data vector x to time-domain sampling windows rather than frequency-domain quadrature components.
[0076] The following provides an example of a detection test for Stage 1.
[0077] The goal of Stage 1 is to decide betweenHo: no target vs Ht: “some target present” G {H1;H2}
[0078] We take the approach of ignoring the distinction between A and B when detecting a target. In other words, we just ask: “Is there any non-zero signal present?”
[0079] Under Gaussian assumptions, the likelihood of x under Hois
[0080] Under Gaussian assumptions with unknown amplitude, the likelihood of x underfor a fixed amplitude aAis
[0081] We may calculate value of the amplitude aAthat maximises the likelihood of x under:maxpaASo that / \ N log maxpCxI#!,^) = — — log\ aA) 2andlogp1x). Thus, the log-likelihood version of the Stage-1Test may be written as2Si(x) = log
[0082] We declare “detection” if Sx(x) > ]1for some threshold ]1set to meet a desired probability of false alarm.
[0083] The following provides an example of a classification test for Stage 2.
[0084] Once Stage 1 declares a target present, we then decide betweenvs. H2. In the standard Neyman-Pearson or Bayesian classification framework, the optimal test for known signals but unknown amplitudes is:maxpS2(x) = logaAmaxpO-BUnder( \ N 1 log maxpCxIf / a = --log(27r) - -log |\ aA) 2 2
[0085] Similarly, for H2. Substituting into S2(x), most common terms cancel, leaving a decision rule of the formDecision:(S2(x) > p2-> " Declare target A",( otherwise -> " Declare target B".where the choice of the threshold p2controls the desired trade-off between incorrectly labelling B as A and incorrectly labelling A as B.
[0086] Thus, Stage 2 uses a ratio of the best-fit, amplitude-matched likelihoods for A and B, weighted by their priors. This is the classical generalised likelihood ratio test (GLRT) classifier for unknown amplitudes in Gaussian noise.
[0087] Crucially, the pre-whitening used in Stage 1 makes the first indicator output insensitive to at least a first component due to the soil. This makes the first indicator output of high signal -to-noise ratio when detecting targets of interest in mineralised soils. Typically, the first indicator output is insensitive to the instantaneous reactive response from the soil, which has a reactive component that is independent of frequency. Typically, the first indicator output is also insensitive to the log-uniform VRM response from the soil, which has a resistive component that is approximately independent of frequency and a reactive component that is a log-linear function of frequency.
[0088] The pre-whitening used in Stage 2 makes the second indicator output insensitive to at least a first component due to the soil. This makes the second indicator output of high signal-to-noise ratio when detecting targets of interest in mineralised soils. Typically, the second indicator output is insensitive to the log-uniform VRM response from the soil, which has a resistive component that is approximately independent of frequency and a reactive component that is a log-linear function of frequency. The second indicator output may be sensitive to a second soil component. For example, the second soil component may be the log-linear VRM response from the soil, which has a resistive component that is a log-linear function of frequency and a reactive component that is a log-quadratic function of frequency.
[0089] With appropriate pre-whitening used in Stage 2, many hot rocks may be discriminated from targets of interest on the basis that the magnitude of instantaneous reactive response from the hot rock is inconsistent with being due to noise from the surrounding soil matrix.
[0090] In conventional metal detectors, the phase-space of responses is typically partitioned into regions associated with targets of interest, ferrous objects, and hot rocks. The instantaneous reactive response is ordinarily employed to distinguish targets of interest from ferrous objects and hot rocks. However, in highly mineralised soils, such as those prevalent in goldfields, the surrounding matrix itself elicits a substantial instantaneous reactive response that may overshadow the signal of a genuine target. As a result, when a true target is encountered, its instantaneous reactive response may be inflated by the contribution from the surrounding soil matrix and thus fall within the phase-space region normally reserved for ferrous objects and hot rocks. This overlap renders conventional discrimination methods unreliable, since any object may appear in that region simply due to the soil matrix contribution. Accordingly, pre-whitening is applied to normalize the response of the surrounding soil matrix. By transforming responses according to the covariance matrix of the soil’s reactive signature, pre -whitening highlights instantaneous reactive responses that deviate from thetypical soil patern, thereby restoring the discriminatory power of the instantaneous reactive response for accurate target identification in mineralised soils.
[0091] Furthermore, when classifying between ferrous and non-ferrous targets in the presence of mineralised soil, it is beneficial for the discriminating signal to be insensitive to at least the loguniform VRM component from the soil.
[0092] Figures 4A to 4C depict an exemplary implementation of the present disclosure. This exemplary implementation does not limit the scope of this disclosure, for example, the gold nugget (the wanted target) and the hot rock (the unwanted object) in this example may be a non-ferrous conductive object (a wanted object) and a ferrous object (the unwanted object) instead.
[0093] Figure 4A depicts the case when a gold nugget 11 is below a coil of the metal detector during operation. When a gold nugget 11 is present, the first indicator output, which in this case is an audible alert, indicates the presence of a signal of some kind. The second indicator output, which in this case is a meter with the needle pointed toward the right side of the range, indicates that the signal is more likely due to a gold nugget than a hot rock.
[0094] Figure 4B depicts the case when a hot rock 13 is below a coil of the metal detector during operation. When a hot rock 13 is present, the first indicator output, which in this case is an audible alert, indicates the presence of a signal of some kind. The second indicator output, which in this case is a meter with the needle pointed toward the left side of the range, indicates that the signal is more likely due to a hot rock than a gold nugget.
[0095] Figure 4C depicts the case when a gold nugget 15 is below a coil of the metal detector during operation, but at a location deeper than the location of the case depicted in Figure 4A. When a deeper nugget 15 is present, the first indicator output, which in this case is an audible alert, more weakly indicates the presence of a signal of some kind. The second indicator output, which in this case is a meter with the needle pointed toward the middle of the range, indicates a more balanced likelihood of a gold nugget being present compared to the likelihood of a hot rock being present.
[0096] Accordingly, the second indicator output assists the user of the metal detector to decide whether there is a high likelihood that the first indicator output signal is responding to a target of interest or a hot rock. This will save time by reducing the rate of false alarms.
[0097] Of course, it is possible to combine the first indicator output and the second indicator output to form a combined output signal. For example, the combined output signal may present a differentsound, or present a different light or display, etc., when there is a target and that the likelihood of the target being a hot rock is high.
[0098] In this specification the terms “ground” and “soil” are used interchangeably. As understood by a person skilled in the art, the terms “ground” and “soil” mean surface regions of earth in which targets may be contained. The surfaces are often solid, may be homogeneous or may be a combination of various soil types, and may contain moisture or water.
[0099] Those of skill in the art would understand that information and signals may be represented using any of a variety of technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips referenced throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.
[0100] Those of skill in the art would further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.
[0101] The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. For a hardware implementation, processing may be implemented within one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. Software modules, also known as computer programs, computer codes, or instructions, may contain a number of source code or object code segments or instructions, and may reside in any computer readable medium such as a RAM memory, flash memory, ROM memory, EPROM memory, registers, hard disk, a removable disk, a CD-ROM, a DVD-ROM or any other form of computer readable medium. In the alternative, the computer readable medium may be integral to the processor. The processor and the computer readable medium may reside in an ASIC or related device. The software codes may be stored in amemory unit and executed by a processor. The memory unit may be implemented within the processor or external to the processor, in which case it may be communicatively coupled to the processor via various means as is known in the art.
[0102] Throughout the specification and the claims that follow, unless the context requires otherwise, the words “comprise” and “include” and variations such as “comprising” and “including” will be understood to imply the inclusion of a stated integer or group of integers, but not the exclusion of any other integer or group of integers.
[0103] The reference to any prior art in this specification is not, and should not be taken as, an acknowledgement of any form of suggestion that such prior art forms part of the common general knowledge.
[0104] It will be appreciated by those skilled in the art that the disclosure is not restricted in its use to the particular application described. Neither is the present disclosure restricted in its preferred embodiment with regard to the particular elements and / or features described or depicted herein. It will be appreciated that the disclosure is not limited to the embodiment or embodiments disclosed, but is capable of numerous rearrangements, modifications and substitutions without departing from the scope of the disclosure as set forth and defined by the following claims.
Claims
1. CLAIMS1. A method for detecting one or more targets of interest in a soil with a metal detector, the method comprising:generating a transmit magnetic field for transmission into the soil based on a transmit signal; receiving a receive magnetic field to induce a receive signal;processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and insensitive to a first component of the receive signal due to the soil;processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and an unwanted second component different from the first component; and producing at least one indicator output signal based on the first signal and the second signal.
2. The method of claim 1, wherein the unwanted second component is due to the soil, ferrous objects, or both.
3. The method of claim 1, wherein the second signal is less sensitive to the one or more targets of interest than the first signal.
4. The method of claim 1, further comprising:producing a first indicator output signal based on the first signal; andproducing a second indicator output signal based on the second signal.
5. The method of claim 4, wherein the first indicator output signal and the second indicator output signal are presented to a user of the metal detector simultaneously.
6. The method of claim 1, wherein the first component comprises an instantaneous reactive magnetic response, or a log-uniform distributed magnetic soil VRM component, or both.
7. The method of claim 1 or 6, wherein the second signal discriminates between the one or more targets of interest and a second component due to the soil, the second component being different from the first component.
8. The method of claim 1, wherein the second signal discriminates between the one or more targets of interest that are non-ferrous and a second component due to ferrous objects.
9. The method of claim 1, wherein the second component comprises a log-linear distributed magnetic soil VRM component, an instantaneous reactive magnetic response, or both.
10. The method of claim 1, wherein the second component comprises a resistive component that increases approximately linearly with frequency.
11. The method of claim 1, wherein the step of processing the receive signal to produce a first signal that is sensitive to the one or more targets of interest and is insensitive to a first component of the receive signal due to the soil comprises one or both of the following: producing a projection of the receive signal that is orthogonal to the first component due to the soil in a time domain or frequency domain; and pre-whitening based on a historical noise measurement.
12. The method of claim 1, wherein the step of processing the receive signal to produce a second signal that is insensitive to the first component of the receive signal due to the soil and that discriminates between the one or more targets of interest and a second component comprises: prewhitening based on a historical noise measurement.
13. The method of claim 1 or 12, wherein the step of producing at least one indicator output signal based on the first signal and the second signal comprises: comparing a likelihood of a first target of interest and noise being present with a likelihood of noise only being present; and comparing a likelihood of the first target of interest being present with a likelihood of the second component being present.
14. The method of claim 4, wherein the first indicator output and the second indicator output are in the form of one or more of visual, audio, or haptic indicators.
15. The method of claim 1, wherein the step of producing at least one indicator output signal based on the first signal and the second signal comprises applying a high-pass or band-pass filter with at least 6dB of attenuation at 0 Hz.
16. The method of claim 5, wherein the first indicator output and the second indicator output are simultaneously presented via a user interface to a user when the metal detector is used in a non-motion mode.
17. The method of claim 1, wherein the one or more targets of interest comprise gold nuggets weighing between 0.01g and 100g each.
18. The method of claim 1, wherein the step of processing the receive signal to produce the second signal comprises combining low, medium, and high frequency components of the receivesignal to produce two channels that are both insensitive to the first component; and wherein the first of the two channels is insensitive to a particular target of interest and the second of the two channels is sensitive to the particular target of interest.
19. The method of claim 1, wherein the metal detector is a handheld metal detector.
20. A metal detector configured to perform the method of claim 1.
