Magnetometer Data Visualised As Colour Images Reveals Hidden Signals

Researchers at the Indian Institute of Technology, led by Manas Pandey, have developed a novel visualisation technique designed to enhance anomaly detection within magnetometer arrays. The method constructs a ‘cross-sensor RGB spectrogram’ by mapping the power spectra derived from three magnetometers into red, green, and blue colour channels, thereby revealing subtle inter-sensor relationships that are frequently overlooked in conventional data analysis procedures. This approach facilitates the swift identification of coherent magnetic activity, potential sensor malfunctions, and asymmetric magnetic sources, and demonstrates compatibility with both established fluxgate sensors and emerging quantum magnetometer technologies, including optically pumped magnetometers, nitrogen-vacancy centre arrays, and SQUIDs. Crucially, the technique operates on scalar data, offering a versatile and self-contained toolkit applicable to any magnetometer monitoring pipeline, and directly addresses the critical challenge of distinguishing quantum-limited noise from technical artefacts inherent in advanced quantum sensing systems.

Rapid anomaly detection using cross-sensor RGB spectrogram visualisation

The cross-sensor RGB spectrogram demonstrably improves the speed of anomaly recognition, reducing visual search time for subtle disturbances by a factor of three when compared to standard magnetometer data analysis techniques. Traditionally, the identification of asymmetric magnetic sources or sensor faults necessitated a laborious and time-consuming comparison of individual sensor power spectra. This new method consolidates this information into a single, easily interpretable image, enabling analysts to immediately recognise anomalies through discernible colour changes. The underlying principle relies on the fact that magnetic anomalies manifest as alterations in the magnetic field, and these alterations will be detected, potentially with varying degrees of intensity and phase, by multiple sensors within an array. Standard analysis often focuses on the magnitude of the signal from each sensor independently, losing information about the relationships between those signals.

The visualisation technique maps the power spectra from three concurrently sampled magnetometers directly into the red, green, and blue channels of an image. Representing inter-sensor coherence, the degree to which the signals from different sensors correlate, as neutral tones (greys and whites), and unique spectral energy as saturated colours, provides an intuitive visual representation of the data. A sensor experiencing a fault, for example, might contribute strongly to one colour channel (red, green, or blue) while showing little correlation with the other two, resulting in a saturated colour patch. Conversely, a coherent, widespread magnetic disturbance affecting all sensors equally would appear as a neutral grey or white, indicating strong agreement between the sensors. The power spectrum itself is calculated using the Fast Fourier Transform (FFT), a standard signal processing technique that decomposes a time-series signal into its constituent frequencies, revealing the energy present at each frequency. The magnitude of the FFT at each frequency is then used to determine the colour intensity for each channel.

This colour-anomaly taxonomy allows for the differentiation of several key phenomena: coherent broadband activity (uniform colour), single-sensor faults (saturated colour), asymmetric pairwise sources (mixed colours indicating differing responses between sensor pairs), and slow temporal drift (gradual colour changes over time). The method’s applicability extends beyond conventional fluxgate sensors to encompass a range of quantum magnetometers, including optically pumped magnetometers (OPMs) which rely on the absorption of light by alkali vapours, nitrogen-vacancy (NV) centre arrays in diamond, and Superconducting Quantum Interference Devices (SQUIDs). This broad compatibility is particularly valuable as these quantum sensors offer increased sensitivity but also introduce new sources of noise and artefacts, making the differentiation of genuine signals from technical issues more challenging. The technique aids in this differentiation by providing a visual indicator of sensor health and coherence. Operating on scalar magnitude time series data, the construction can be seamlessly integrated into any monitoring pipeline utilising a synchronously sampled magnetometer triad, and a long-window variant of the spectrogram is capable of resolving features in the ultra-low frequency band, crucial for geomagnetic studies.

Cross-sensor spectrograms enable faster assessment of magnetometer array performance

Geomagnetic observatories, dedicated to the precise measurement of the Earth’s magnetic field, and laboratory setups requiring high-precision magnetic field control, are increasingly reliant on arrays of magnetometers to achieve the necessary levels of accuracy and spatial resolution. Current analysis methods typically involve a meticulous, individual examination of each sensor’s data, a process that can be both time-consuming and prone to overlooking subtle, yet significant, inter-sensor relationships. This visualisation technique offers a rapid, holistic overview of array health and activity, providing a significant advantage in real-time monitoring and data quality control. Functioning as a methodological building block, further development is needed to fully exploit the speed gains offered by the spectrogram, potentially through automated anomaly detection algorithms. By presenting a comprehensive view of the array’s performance, the method simplifies the identification of faults and disturbances, functioning effectively with both traditional fluxgate sensors and advanced quantum magnetometers such as SQUIDs, which are renowned for their exceptional sensitivity but require careful calibration and noise characterisation. Neutral colours within the spectrogram indicate strong agreement between sensors, suggesting a consistent and reliable measurement, while saturated hues highlight unique energy signatures, revealing relationships previously obscured by individual sensor analysis. This is achieved by mapping the power spectra into the red, green, and blue channels, providing a concise and intuitive visual representation of inter-sensor coherence and spectral energy distribution. The technique’s ability to quickly identify deviations from expected behaviour is particularly valuable in environments where rapid response to magnetic disturbances is critical, such as space weather monitoring or the detection of magnetic anomalies associated with geological activity.

Researchers at the Indian Institute of Technology, led by Manas Pandey, have developed a novel visualisation technique designed to enhance anomaly detection within magnetometer arrays. The method constructs a ‘cross-sensor RGB spectrogram’ by mapping the power spectra derived from three magnetometers into red, green, and blue colour channels, thereby revealing subtle inter-sensor relationships that are frequently overlooked in conventional data analysis procedures. This approach facilitates the swift identification of coherent magnetic activity, potential sensor malfunctions, and asymmetric magnetic sources. It also demonstrates compatibility with both established fluxgate sensors and emerging quantum magnetometer technologies, including optically pumped magnetometers, nitrogen-vacancy centre arrays, and SQUIDs. Crucially, the technique operates on scalar data, offering a versatile and self-contained toolkit applicable to any magnetometer monitoring pipeline, and directly addresses the critical challenge of distinguishing quantum-limited noise from technical artefacts inherent in advanced quantum sensing systems.

Rapid anomaly detection using cross-sensor RGB spectrogram visualisation

The cross-sensor RGB spectrogram demonstrably improves the speed of anomaly recognition, reducing visual search time for subtle disturbances by a factor of three when compared to standard magnetometer data analysis techniques. Traditionally, the identification of asymmetric magnetic sources or sensor faults necessitated a laborious and time-consuming comparison of individual sensor power spectra. This new method consolidates this information into a single, easily interpretable image, enabling analysts to immediately recognise anomalies through discernible colour changes. The underlying principle relies on the fact that magnetic anomalies manifest as alterations in the magnetic field, and these alterations will be detected, potentially with varying degrees of intensity and phase, by multiple sensors within an array. Standard analysis often focuses on the magnitude of the signal from each sensor independently, losing information about the relationships between those signals.

The visualisation technique maps the power spectra from three concurrently sampled magnetometers directly into the red, green, and blue channels of an image. Representing inter-sensor coherence, the degree to which the signals from different sensors correlate, as neutral tones (greys and whites), and unique spectral energy as saturated colours, provides an intuitive visual representation of the data. A sensor experiencing a fault, for example, might contribute strongly to one colour channel (red, green, or blue) while showing little correlation with the other two, resulting in a saturated colour patch. Conversely, a coherent, widespread magnetic disturbance affecting all sensors equally would appear as a neutral grey or white, indicating strong agreement between the sensors. The power spectrum itself is calculated using the Fast Fourier Transform (FFT), a standard signal processing technique that decomposes a time-series signal into its constituent frequencies, revealing the energy present at each frequency. The magnitude of the FFT at each frequency is then used to determine the colour intensity for each channel.

This colour-anomaly taxonomy allows for the differentiation of several key phenomena: coherent broadband activity (uniform colour), single-sensor faults (saturated colour), asymmetric pairwise sources (mixed colours indicating differing responses between sensor pairs), and slow temporal drift (gradual colour changes over time). The method’s applicability extends beyond conventional fluxgate sensors to encompass a range of quantum magnetometers, including optically pumped magnetometers (OPMs) which rely on the absorption of light by alkali vapours, nitrogen-vacancy (NV) centre arrays in diamond, and Superconducting Quantum Interference Devices (SQUIDs). This broad compatibility is particularly valuable as these quantum sensors offer increased sensitivity but also introduce new sources of noise and artefacts, making the differentiation of genuine signals from technical issues more challenging. The technique aids in this differentiation by providing a visual indicator of sensor health and coherence. Operating on scalar magnitude time series data, the construction can be seamlessly integrated into any monitoring pipeline utilising a synchronously sampled magnetometer triad, and a long-window variant of the spectrogram is capable of resolving features in the ultra-low frequency band, crucial for geomagnetic studies.

Cross-sensor spectrograms enable faster assessment of magnetometer array performance

Geomagnetic observatories, dedicated to the precise measurement of the Earth’s magnetic field, and laboratory setups requiring high-precision magnetic field control, are increasingly reliant on arrays of magnetometers to achieve the necessary levels of accuracy and spatial resolution. Current analysis methods typically involve a meticulous, individual examination of each sensor’s data, a process that can be both time-consuming and prone to overlooking subtle, yet significant, inter-sensor relationships. This visualisation technique offers a rapid, holistic overview of array health and activity, providing a significant advantage in real-time monitoring and data quality control. Functioning as a methodological building block, further development is needed to fully exploit the speed gains offered by the spectrogram, potentially through automated anomaly detection algorithms. By presenting a comprehensive view of the array’s performance, the method simplifies the identification of faults and disturbances, functioning effectively with both traditional fluxgate sensors and advanced quantum magnetometers such as SQUIDs, which are renowned for their exceptional sensitivity but require careful calibration and noise characterisation. Neutral colours within the spectrogram indicate strong agreement between sensors, suggesting a consistent and reliable measurement, while saturated hues highlight unique energy signatures, revealing relationships previously obscured by individual sensor analysis. This is achieved by mapping the power spectra into the red, green, and blue channels, providing a concise and intuitive visual representation of inter-sensor coherence and spectral energy distribution. The technique’s ability to quickly identify deviations from expected behaviour is particularly valuable in environments where rapid response to magnetic disturbances is critical, such as space weather monitoring or the detection of magnetic anomalies associated with geological activity.

The researchers developed a new visualisation technique called the cross-sensor RGB spectrogram to analyse data from arrays of three magnetometers. This method transforms power spectra into a single image, revealing relationships between sensors that are often missed by standard analysis. The resulting image uses colour to indicate the coherence of signals, with neutral shades representing agreement and saturated colours highlighting unique activity or potential faults. This approach provides a rapid overview of array health and simplifies the identification of disturbances, functioning with both fluxgate sensors and SQUIDs. The authors suggest future work could focus on automating anomaly detection using this technique.