A method for positioning and alarming hidden camera equipment combined with ambient light perception

Abstract

The present invention relates to the field of surreptitious photography identification technology, and discloses a method for locating and alerting surreptitious photography devices in combination with ambient light perception. The method comprises: deploying a sensor array, signal receiving nodes, and image acquisition devices in a target area to acquire real-time ambient light intensity and color distribution change data and generate its change trend; using statistical analysis to identify abnormal ambient light changes and output the coordinates and features of the abnormal light changes; acquiring and preprocessing corresponding images, extracting surreptitious photography device features using image algorithms, and verifying them in combination with a rule base to determine candidate locations; then obtaining precise location coordinates through triangulation; if both ambient light anomalies and surreptitious photography features are present at the coordinates, the device is determined to be a surreptitious photography device, and an alarm system receives the coordinates and triggers an alarm and interference measures. The present invention improves the surreptitious photography identification accuracy and positioning capabilities in complex scenarios.

Description

Technical Field

[0001] The present invention relates to the field of hidden camera identification technology, and in particular to a hidden camera device positioning and alarm method combined with ambient light perception. Background Art

[0002] With the widespread use of smart devices, hidden cameras are becoming increasingly concealed and smaller. They are equipped with wireless transmission and low-power signal transmission features, making them easily disguised in common environmental objects such as sockets, mirrors, and smoke alarms, posing a serious threat to public privacy and security. Existing hidden camera detection methods primarily rely on image analysis or wireless signal detection, but these often suffer from the following technical limitations:

[0003] Image processing methods rely on a clear field of vision and are sensitive to changes in ambient lighting. They are prone to false detection or missed detection in weak or complex lighting conditions. Wireless signal detection methods cannot accurately locate the target device, especially when the target device has weak power or there is signal interference. It is difficult to effectively identify the specific location of the device. A single perception method has poor adaptability and is difficult to cope with the complex environment and diverse disguise threats of voyeurism in real scenarios.

[0004] Therefore, there is an urgent need for a hidden camera detection and prevention method that integrates multi-source information, has environmental adaptability, and can achieve fast and accurate positioning and real-time alarms, so as to improve the security capabilities of public places and private spaces. Summary of the Invention

[0005] In view of this, the present invention proposes a method for positioning and alarming hidden camera devices combined with ambient light perception, aiming to solve the problems in current technology of low recognition rate of hidden camera behavior in complex scenes and lack of ability to accurately locate hidden camera devices.

[0006] The present invention proposes a method for positioning and alarming a hidden camera device in combination with ambient light sensing, comprising:

[0007] Deploy sensor arrays, signal receiving nodes, and image acquisition devices in the target area;

[0008] Obtain real-time ambient light intensity changes and ambient color distribution changes to generate ambient light intensity change trend data and ambient color distribution change trend data;

[0009] The ambient light intensity change trend data and the ambient color distribution change trend data are statistically analyzed to identify abnormal ambient light changes, and the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change are output;

[0010] Acquire an image at the coordinate position of the abnormal ambient light change, preprocess the image to obtain a preprocessed image, and use an image processing algorithm to extract features from the preprocessed image to obtain features of the hidden camera device;

[0011] The rule base verification method is used to verify the validity of the hidden camera device features, and the candidate location and feature description of the hidden camera device are obtained;

[0012] Use triangulation method to locate the candidate location of the hidden camera device and obtain the precise location coordinates of the hidden camera device;

[0013] If the characteristic description of abnormal ambient light changes and the characteristic description of the hidden camera device exist at the precise location coordinates of the hidden camera device at the same time, the hidden camera device is determined to be real, and the precise location of the hidden camera device is output to the alarm system, which will issue an alarm and take interference measures.

[0014] Furthermore, the ambient light intensity change trend data and the ambient color distribution change trend data are subjected to statistical analysis methods to identify abnormal ambient light changes, and the output is the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change. The specific content is: time series modeling and multidimensional statistical analysis are used for the collected ambient light intensity change trend data and the ambient color distribution change trend data, firstly, a light intensity change curve and a corresponding color distribution histogram in frames are constructed, and the local mean, variance, and skewness statistical features are extracted using the sliding window technology; then, the threshold discrimination method and the principal component analysis method are introduced to identify data points in the local area that deviate significantly from the overall trend, and potential abnormal areas are preliminarily screened out. The abnormal area is further determined in specific location by combining the spatial coordinate mapping technology, and the characteristic information of the average brightness, hue change rate, and saturation mutation degree of the abnormal area is extracted to construct a multidimensional feature vector of the abnormal ambient light change, namely, the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change.

[0015] Furthermore, the description of the abnormal ambient light change characteristics includes but is not limited to local light intensity abnormality, abnormality of specific wavelength light, and sudden change of ambient color temperature.

[0016] Furthermore, the method acquires an image at the coordinate position where the abnormal ambient light changes, preprocesses the image to obtain a preprocessed image, and uses an image processing algorithm to perform feature extraction on the preprocessed image to obtain the characteristics of the hidden camera device. The specific contents are as follows: first, preprocessing operations are performed on the acquired image, including image enhancement, denoising, contrast enhancement, edge sharpening, and region of interest extraction, to obtain a preprocessed image; then, a multi-scale edge detection algorithm is used in combination with Harris corner detection on the preprocessed image to extract the edge features and structural contours of the potential hidden camera device, and reflected light spot detection and circular / point highlight area extraction algorithms are applied to the image to identify lens reflections, highlight points or geometric features that appear under abnormal lighting conditions; combined with morphological analysis, areas with typical "lens hole" features are further screened; finally, the hidden camera characteristics of the device are obtained by constructing a feature vector and introducing a trained target detection model for classification and judgment.

[0017] Furthermore, the region of the “lens hole” feature includes, but is not limited to, a circular structure, strong edge regularity, and a sudden change in central reflectivity;

[0018] The feature vector includes structural texture, spot size, brightness gradient, shape descriptor, and spectral reflectance characteristics.

[0019] Furthermore, the rule base verification method is used to verify the validity of the hidden camera device features to obtain the candidate location of the hidden camera device and the hidden camera device feature description. The specific content is: the hidden camera device features are normalized, and then the similarity is calculated with the corresponding device template features in the pre-built hidden camera device feature rule base, and multiple judgment rules are set. If the hidden camera device features have matching results under multiple rules, the candidate location of the hidden camera device and the hidden camera device feature description are output.

[0020] Furthermore, the triangulation positioning method is used for the candidate position of the hidden camera device to obtain the precise position coordinates of the hidden camera device. The specific content is: the signal receiving node located at the candidate position of the hidden camera device receives the electromagnetic signal strength in the direction of the hidden camera device, and according to the RSSI and distance mapping model, the signal strength is converted into a corresponding distance estimation value, and multiple circular intersection areas are constructed with the signal receiving node as the center and the distance estimation value as the radius. The optimal overlapping point of the cross-coverage area of multiple nodes is calculated by the minimum mean square error method, which is the preliminary positioning point of the hidden camera device. The preliminary positioning point of the hidden camera device is combined with the vertical distribution data of the receiving antenna to further obtain the three-dimensional coordinates of the hidden camera device in space, that is, the precise position coordinates of the hidden camera device.

[0021] Furthermore, the signal receiving node located at the candidate location of the hidden camera device receives the electromagnetic signal strength in the direction of the hidden camera device and converts the signal strength into a corresponding distance estimation value according to the RSSI and distance mapping model, which is expressed as:

[0022] ;

[0023] in, is the estimated distance, To calibrate the signal strength value on site, The electromagnetic signal strength in the direction of the hidden camera device. is the path loss factor.

[0024] Furthermore, multiple circular intersection areas are constructed with the signal receiving node as the center and the distance estimation value as the radius. The optimal overlapping point of the cross-coverage areas of multiple nodes is calculated by the minimum mean square error method, which is the preliminary positioning point of the hidden camera device and is expressed as:

[0025] ;

[0026] in, This is the initial positioning point for the hidden camera device. is the total number of signal receiving nodes, is the first value estimated by the RSSI mapping model i The distance between the receiving node and the target device, For the i The coordinates of the signal receiving node.

[0027] Furthermore, if the characteristic description of abnormal ambient light changes and the characteristic description of the hidden camera device exist at the precise position coordinates of the hidden camera device at the same time, the specific content of determining that the hidden camera device is real is: comparing and analyzing the ambient light change characteristics and the hidden camera device characteristics. If the ambient light change characteristics and the hidden camera device characteristics appear at the same coordinate position at the same time, the hidden camera device is determined to be real based on spatial consistency and feature superposition.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] The present invention relates to a method for positioning and alarming a hidden camera device in combination with ambient light perception. The method uses the changes in ambient light intensity and color distribution trends as signal sources for precursors to abnormal behavior, extracts potential abnormal light change areas through statistical analysis and time series modeling, and combines image feature recognition and radio frequency positioning technology to achieve fusion judgment from environmental signals, image information and electromagnetic signals, thereby greatly improving the robustness and accuracy of hidden camera device recognition. The present invention introduces an automatic modeling and abnormal discrimination mechanism for ambient light change trends, which can quickly lock highly suspicious areas in a wide range of scenes, avoid blind scanning and a large amount of invalid calculations, effectively improve detection efficiency, and reduce false positives and missed reports. The present invention combines multiple algorithms such as image preprocessing, edge detection and light spot analysis. , established a morphological feature model of the "lens hole" of the hidden camera device, further improving the interpretability and recognition success rate of the device identification; the present invention enhances the robustness and consistency of feature recognition by matching and verifying the extracted hidden camera device features with the pre-built rule library, and significantly reduces the risk of misidentification caused by misjudgment of a single algorithm; the present invention combines RSSI signal strength and triangulation positioning method for spatial positioning, accurately calculates the three-dimensional coordinates of the hidden camera device under multi-node deployment, and supports high-precision positioning and multi-dimensional position reconstruction; the "feature superposition consistency judgment mechanism" proposed in the present invention combines light anomalies and physical device features in space to ensure that the alarm device has a high degree of credibility and avoids false alarms or repeated alarms caused by a single feature. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0031] Figure 1 This is a flow chart of a method for positioning and alarming a hidden camera device combined with ambient light perception according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0033] See Figure 1As shown, an embodiment of the present invention provides a method for positioning and alarming a hidden camera device in combination with ambient light perception, including:

[0034] S1: Deploy sensor arrays, signal receiving nodes, and image acquisition devices in the target area;

[0035] S2: Obtain real-time ambient light intensity changes and ambient color distribution changes to generate ambient light intensity change trend data and ambient color distribution change trend data;

[0036] S3: Using a statistical analysis method to analyze the ambient light intensity change trend data and the ambient color distribution change trend data, abnormal ambient light changes are identified, and the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change are output;

[0037] S4: Acquire an image at the coordinate position of the abnormal ambient light change, preprocess the image to obtain a preprocessed image, and use an image processing algorithm to extract features from the preprocessed image to obtain features of the hidden camera device;

[0038] S5: Using a rule base verification method to verify the validity of the hidden camera device features, and obtaining a candidate location of the hidden camera device and a description of the hidden camera device features;

[0039] S6: Using triangulation method to locate the candidate location of the hidden camera device to obtain the precise location coordinates of the hidden camera device;

[0040] S7: If the characteristic description of abnormal ambient light changes and the characteristic description of the hidden camera device exist at the precise location coordinates of the hidden camera device, the hidden camera device is determined to be real, and the precise location of the hidden camera device is output to the alarm system, which issues an alarm and takes interference measures.

[0041] Furthermore, sensor arrays, signal receiving nodes and image acquisition devices are deployed in the target area. Specifically, RGB sensors and infrared sensors are evenly deployed in the target area to collect parameter information such as ambient light intensity, spectral distribution, and color temperature changes in real time, build a continuous illumination model and identify abnormal light change events. A dense sampling network is formed between multiple sensors to facilitate the analysis of local light mutation phenomena.

[0042] Furthermore, the system captures real-time changes in ambient light intensity and color distribution, generating trend data for both. Ambient light intensity fluctuations reflect the illumination fluctuations in the target area at different points in time. The lens or screen of a hidden camera device, due to its reflective properties, material, or surface coating, may cause short-term aberrations in light intensity or stroboscopic signals in localized areas, resulting in a trend inconsistent with the background illumination. Ambient color distribution changes are used to monitor the dynamic adjustment of color components across the region. The optical structure of a hidden camera device (such as filters, lens refraction, and infrared filter elements) may enhance or suppress reflections of specific wavelengths, causing localized color shifts or spectral anomalies. Through continuous analysis of trend data, the system can model and characterize aberrant optical behavior over time. This allows for early warning and risk assessment of potential activity without requiring direct identification of the hidden camera device's physical form.

[0043] It's important to note that by generating real-time data on changing trends in ambient light intensity and color distribution, the system establishes a stable background optical model and, using a dynamic deviation recognition mechanism, effectively captures the subtle optical perturbations caused by hidden camera devices. This approach improves the robustness of the hidden camera detection system in complex lighting environments and is particularly suitable for concealed spaces, statically controlled areas, or scenarios with multiple light sources, enabling intelligent, non-invasive detection of hidden camera risks.

[0044] Furthermore, the ambient light intensity change trend data and the ambient color distribution change trend data are subjected to statistical analysis methods to identify abnormal ambient light changes, and the output is the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change. The specific contents are as follows: time series modeling and multidimensional statistical analysis are used for the collected ambient light intensity change trend data and the ambient color distribution change trend data. First, a light intensity change curve and a corresponding color distribution histogram in frames are constructed, and the sliding window technology is used to extract the local mean, variance, and skewness statistical features; then the threshold discrimination method and the principal component analysis method are introduced to identify data points in the local area that deviate significantly from the overall trend, and potential abnormal areas are preliminarily screened out. The abnormal area is further determined by combining the spatial coordinate mapping technology to determine the specific location, and the average brightness, hue change rate, and saturation mutation degree characteristic information of the abnormal area are extracted to construct a multidimensional feature vector of the abnormal ambient light change, namely the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change.

[0045] Specifically, the collected data on ambient light intensity and color distribution trends are processed using time series modeling and multidimensional statistical analysis. First, using frames as the basic time unit, the average brightness value of each region within each frame is extracted and recorded to construct a continuously changing ambient light intensity curve. This curve not only reflects the overall temporal trend of illumination but also can be used to identify stability, periodicity, or suddenness of illumination changes. Simultaneously, data from the RGB channels and, where possible, infrared or other specific bands, is extracted from each frame, and a color distribution histogram is calculated to form a statistical representation of the ambient color characteristics of the current frame. The color distribution histogram can capture information such as color shifts, saturation fluctuations, and dominant hue shifts, providing a data foundation for subsequent color trend modeling. A sliding time window of fixed length (e.g., 3, 5, or 10 frames) is set on each time series. Within each window, local statistical characteristics for that time period are calculated, including metrics such as mean, variance, and skewness. The local mean reflects the average level of overall brightness or color distribution within a time slice, the variance measures the fluctuations in brightness or color, and the skewness reveals the asymmetry of the data distribution and is often used to capture extreme trends in a particular direction. The sliding window moves continuously along the time dimension, achieving full sequence coverage and scanning, thereby forming a continuous perception of local trends in illumination and color data. For color distribution histogram processing, statistics can be extracted by channel, or the overall color histogram can be normalized and its trend extracted. If there is a sharp rise or fall in light intensity within a short period of time, or a sudden shift in the main color peak, the corresponding local variance and skewness values will fluctuate significantly, providing preliminary criteria for anomaly detection. A suitable threshold range is set for the local mean, variance, skewness and other statistical feature values of each frame. If the deviation from the overall trend value (for example, the set benchmark statistics, such as the global mean or standard deviation) exceeds the predetermined threshold, the data point is judged to be a potential anomaly. Subsequently, principal component analysis (PCA) is used to reduce the dimensionality of the multidimensional feature vector, extract the principal components, analyze the weight of each feature's contribution to the anomaly point, and further identify data points that show abnormal trends in the low-dimensional space. PCA decomposes the covariance matrix of multidimensional features such as light intensity and color distribution into eigenvalues, maps the original features to the new space, and highlights the key change trends of various features. After PCA-based dimensionality reduction, it can more accurately capture complex lighting changes and color anomalies, reduce noise interference, and avoid misjudging anomalies due to local feature fluctuations; finally, the coordinate information of the abnormal area is fused with the sensor data through image processing algorithms, and the spatial coordinate mapping method is used to associate the abnormal area in the two-dimensional image with the physical location in the actual scene.Spatial coordinate mapping technology is based on the spatial layout and relative position of each device in the sensor array. It uses the timestamp and spatial position information of the sensor data to generate the coordinate position of abnormal ambient light changes that match the physical space, thereby accurately marking the precise location where the abnormal light change occurs. After positioning is completed, key features such as average brightness, hue change rate, and saturation mutation degree are further extracted for each abnormal ambient light change coordinate position.

[0046] It should be noted that light intensity curve analysis can be used to determine whether a scene is in a state of continuous and uniform illumination, or whether there are potential disturbances such as intermittent strong light or rapidly moving shadows. Due to mirror reflections, filtering differences, and other factors, hidden camera equipment can cause significant deviations in the statistical characteristics of brightness or color in specific frames. The constructed time series curves and color histograms help to detect these subtle but regular differences. Sliding window statistics can effectively capture local sudden changes and are particularly sensitive to short-term brightness / color fluctuations caused by lens reflections and transient fill light changes in hidden camera equipment. By combining statistical analysis with dimensionality reduction methods, potential anomalies can be intelligently identified in a data-driven manner without the need to manually set excessive threshold conditions or rely on human judgment, thereby improving the system's automation level. Through comprehensive analysis of multiple features, true anomalies can be more effectively identified, avoiding misidentification or missed detection, and improving the system's stability and accuracy. For complex environmental changes (such as reflected light and sensor errors), unnecessary noise can be filtered out, improving the accuracy of anomaly identification.

[0047] Furthermore, the characteristics of abnormal ambient light changes include local light intensity anomalies, anomalies of specific wavelengths of light, sudden changes in ambient color temperature, non-uniformity of light distribution, distortion of spectral components, and the amplitude of light intensity fluctuations caused by light source fluctuations.

[0048] Specifically, localized light intensity anomalies can be assessed by monitoring brightness fluctuations in a specific area. Under normal circumstances, light intensity should be relatively stable. A sudden increase or decrease in localized light intensity may indicate equipment failure, reflection, or sudden interference from an external light source. Abnormalities in specific wavelengths of light refer to abnormal changes in the light intensity of a certain wavelength (such as red or blue light), which typically reflect abnormalities in certain physical or chemical processes, such as changes in the emission spectrum of the light source or changes in the characteristics of the reflective surface. Sudden changes in ambient color temperature refer to rapid and significant changes in the color temperature of the light source (such as from cool to warm tones). This change may be caused by a malfunction in the ambient lighting system, sensor interference, or a change in the light source type.

[0049] It's important to note that by comprehensively considering multiple characteristics (light intensity, wavelength, color temperature, etc.), the system can identify problems at a more granular level. For example, abnormal light intensity may be due to reflection, while abnormal wavelength may be related to sensor failure, and changes in color temperature may indicate a change in the light source. This multi-dimensional analysis helps diagnose problems more quickly and accurately, and locate the specific light source or device.

[0050] Furthermore, an image at the coordinate position of the abnormal ambient light change is obtained, the image is preprocessed to obtain a preprocessed image, and the image processing algorithm is used to extract features from the preprocessed image to obtain the characteristics of the hidden camera device. The specific contents are as follows: first, the obtained image is preprocessed, including image enhancement, denoising, contrast enhancement, edge sharpening and region of interest extraction, to obtain a preprocessed image. Subsequently, the preprocessed image is subjected to a multi-scale edge detection algorithm combined with Harris corner detection to extract the edge features and structural contours of the potential hidden camera device, and the image is subjected to reflected light spot detection and circular / point highlight area extraction algorithms to identify lens reflections, highlight points or geometric features that appear under abnormal lighting conditions. Combined with morphological analysis, the area with typical "lens hole" features is further screened. Finally, the device hidden camera characteristics are obtained by constructing a feature vector and introducing a trained target detection model for classification and judgment.

[0051] Specifically, the acquired image is adjusted for overall brightness and contrast using methods such as histogram equalization and adaptive enhancement (such as CLAHE) to improve the visibility and detail expression of the image under different lighting conditions. Gaussian filtering, median filtering, wavelet denoising, and other methods are then used to suppress random noise introduced by sensor thermal noise or environmental interference in the image, while preserving the image structural information as much as possible. Local contrast enhancement algorithms (such as Retinex) are then used for low-contrast images to optimize the dynamic range of the image to better distinguish foreground objects from the background. Edge detection and sharpening algorithms such as the Laplacian operator and Sobel gradient enhancement are then used to improve the clarity and continuity of the edge contours of the target in the image, which is beneficial for subsequent contour extraction and morphological analysis. Finally, color features, texture information, or prior rules are combined to identify areas in the image where hidden cameras or reflective objects may exist, and these areas are intercepted as key processing targets to reduce the computational burden and improve detection accuracy, thereby obtaining a preprocessed image.

[0052] The pre-processed image is then subjected to multi-scale edge detection algorithms (such as Canny, LoG, and DoG) in conjunction with Harris corner detection technology. This first involves multi-layered edge feature extraction and key structural contour location. Multi-scale edge detection uses filter kernels of varying scales to simultaneously capture both fine structures (such as the edge of a pinhole lens) and macroscopic contours (such as the device frame). Harris corner detection can be used to identify areas in the image with sudden angle changes or significant local grayscale variations, thereby extracting the rigid structural features of potential devices.

[0053] Subsequently, algorithms for detecting reflected light spots and extracting circular / point-shaped highlight regions are applied to the image. Specifically, these algorithms utilize methods such as brightness threshold segmentation and shape discrimination (e.g., Hough circle transform) to identify lens reflections, tiny convex reflection points, or unnatural highlight regions that may form under specific lighting angles. These regions often appear as small, highly reflective, nearly circular or elliptical highlights in real-world scenes, serving as important visual signals for hidden camera devices.

[0054] Combined with morphological analysis methods (such as dilation, erosion, opening and closing operations, connected region analysis, etc.), the above extraction results are subjected to structural filtering and boundary reconstruction, from which areas that meet the typical "lens hole" characteristics in terms of geometric shape, symmetry, brightness gradient, etc. are screened out.

[0055] After completing the above feature extraction, we further construct a multidimensional image feature vector, including indicators such as edge contour density, circularity, reflective area brightness distribution, corner point concentration, and local texture uniformity. These features are ultimately fed into a trained object detection model (such as YOLO, Faster R-CNN, and EfficientDet) for image classification and target location determination, resulting in device hidden camera features.

[0056] It should be noted that the enhancement and denoising operations significantly improve the quality degradation problem in the original image caused by low illumination, sensor interference or compression, making the image clearer and more layered, which is convenient for subsequent processing modules to analyze the image content; contrast enhancement and edge sharpening make the difference between the target area and the background more obvious, which helps to improve the accuracy of tasks such as object detection, segmentation, and recognition, especially in low light or complex background conditions; through image denoising and ROI extraction, background areas and noise interference that are not related to the detection target can be effectively filtered out, reducing the impact of invalid information on the system and reducing the probability of misidentification; the extraction of the region of interest significantly reduces the detection time. The detection range enables the system to focus on high-risk areas under limited computing resources, improve overall processing efficiency and achieve rapid response; through multi-scale edge detection and corner recognition, it can accurately extract tiny structures in complex images and enhance the system's recognition ability for covert hidden cameras (such as pinhole lenses and miniature cameras); reflected light spot detection and highlight area extraction are specifically designed to deal with the physical optical phenomena that are easily formed by hidden cameras under specific lighting, which helps to capture reflection features that are difficult to distinguish with the naked eye but have regularity; morphological processing of the detected candidate areas helps to eliminate pseudo-target areas with irregular or unstable shapes, and improve the discrimination accuracy of subsequent recognition models.

[0057] Furthermore, areas with “lens hole” characteristics include, but are not limited to, circular structures, strong edge regularity, and sudden changes in central reflectivity;

[0058] The feature vector includes structural texture, spot size, brightness gradient, shape descriptor, and spectral reflectance characteristics.

[0059] Specifically, the circular structure is determined by the geometric shape of the lens design, the edge regularity is derived from the boundary accuracy of industrial manufacturing, and the central reflectivity mutation is due to the high reflectivity characteristics of the lens glass facing the light source. These are the inevitable physical manifestations of the optical window of the hidden camera device; structural texture features: extract the gray level co-occurrence matrix (GLCM), local binary pattern (LBP) and other texture descriptors of the local area to characterize the characteristics of the relatively smooth surface and low texture complexity of the lens hole area; spot size and distribution: detect the reflected spot area through brightness threshold segmentation, and count its area, shape compactness, and distribution density. Indicators such as Hu moment, Zernike moment, circularity (roundness), aspect ratio, etc. are used to describe the geometric shape of the candidate area to determine whether it is close to the industrial standard circular design. Spectral reflectance characteristics: The reflection intensity and wavelength response of the area in the RGB and infrared channels are analyzed. Some hidden camera devices may have abnormal reflections in specific bands due to coatings or filters (such as enhanced blue light and strong IR reflection).

[0060] It should be noted that through multi-dimensional feature cross-validation (such as structure + brightness + spectrum), misidentification caused by background highlights, non-lens reflections, etc. can be effectively eliminated, and the false alarm rate can be reduced. Multi-level features such as structure, shape, optics, and spectrum enable the system to not only identify conventional hidden cameras, but also adapt to variant devices in non-traditional forms (such as disguised as buttons, screws, etc.). The above features can be used as input dimensions of supervised learning models, and can also be used for target similarity measurement in unsupervised clustering or anomaly detection algorithms, which helps to continuously optimize detection strategies.

[0061] Furthermore, the rule base verification method is used to verify the validity of the hidden camera device features, and the candidate location of the hidden camera device and the hidden camera device feature description are obtained. The specific contents are as follows: the hidden camera device features are normalized, and then the similarity is calculated with the corresponding device template features in the pre-built hidden camera device feature rule base, and multiple judgment rules are set. If the hidden camera device features have matching results under multiple rules, the candidate location of the hidden camera device and the hidden camera device feature description are output.

[0062] Furthermore, the triangulation positioning method is used for the candidate position of the hidden camera device to obtain the precise position coordinates of the hidden camera device. The specific content is: the signal receiving node located at the candidate position of the hidden camera device receives the electromagnetic signal strength in the direction of the hidden camera device, and according to the RSSI and distance mapping model, the signal strength is converted into the corresponding distance estimation value, and multiple circular intersection areas are constructed with the signal receiving node as the center and the distance estimation value as the radius. The optimal overlapping point of the cross-coverage area of multiple nodes is calculated by the minimum mean square error method, which is the preliminary positioning point of the hidden camera device. The preliminary positioning point of the hidden camera device is combined with the vertical distribution data of the receiving antenna to further obtain the three-dimensional coordinates of the hidden camera device in space, that is, the precise position coordinates of the hidden camera device.

[0063] It should be noted that by converting the RSSI values obtained by multiple signal receiving nodes into distances and using geometric construction to perform circular intersection positioning, the initial position of the hidden camera device can be determined from a two-dimensional plane; combined with the distribution data of the receiving nodes at different heights, the three-dimensional coordinates of the hidden camera device can be restored, so that we not only know "where" it is, but also "how high" it is, achieving precise spatial detection. The least mean squares method (Least Mean Squares) is used to fit and solve the intersection points of multiple circular areas, effectively reducing the positioning error caused by RSSI fluctuations, signal obstruction, environmental multipath and other factors, and improving the robustness and accuracy of candidate target points. Image detection is greatly restricted by occlusion, light blocking, angle, etc. Combined with the RSSI wireless signal inversion results, the authenticity of the suspected "lens hole" in the image can be verified from another modality, improving the fault tolerance and judgment credibility of the overall system.

[0064] Furthermore, the signal receiving node located at the candidate location of the hidden camera device receives the electromagnetic signal strength in the direction of the hidden camera device and converts the signal strength into a corresponding distance estimation value according to the RSSI and distance mapping model, which is expressed as:

[0065] ;

[0066] in, is the estimated distance, To calibrate the signal strength value on site, The electromagnetic signal strength in the direction of the hidden camera device. is the path loss factor.

[0067] Furthermore, multiple circular intersection areas are constructed with the signal receiving node as the center and the distance estimation value as the radius. The optimal overlapping point of the cross-coverage area of multiple nodes is calculated by the minimum mean square error method, which is the preliminary positioning point of the hidden camera device and is expressed as:

[0068] ;

[0069] in, This is the initial positioning point for the hidden camera device. is the total number of signal receiving nodes, is the distance between the i-th receiving node and the target device estimated by the RSSI mapping model, is the coordinate of the i-th signal receiving node.

[0070] Furthermore, if the characteristic description of abnormal ambient light changes and the characteristic description of the hidden camera device exist at the precise location coordinates of the hidden camera device at the same time, the specific content of determining whether the hidden camera device is real is: comparing and analyzing the ambient light change characteristics and the hidden camera device characteristics. If the ambient light change characteristics and the hidden camera device characteristics appear at the same coordinate position at the same time, the hidden camera device is determined to be real based on spatial consistency and feature superposition.

[0071] The specific steps involved in outputting the precise location of the hidden camera to the alarm system and initiating the interference measures include: The precise coordinates of the hidden camera's location are transmitted to the alarm system in real time, triggering a multi-level alarm response mechanism. Upon receiving the precise location, the alarm system immediately issues an audible and visual warning signal through the control center, simultaneously initiating a coordinated marking of the coordinate area within the area with the regional monitoring system. Furthermore, a wireless interference module is used to implement directional electromagnetic interference or infrared laser interference in the spatial area where the hidden camera is located, disrupting its image acquisition and data transmission capabilities, effectively blocking the hidden camera's activity and enabling proactive countermeasures.

[0072] It should be noted that the system can lock the alarm to a specific physical location (such as table corners, wall gaps, and ceilings) based on three-dimensional coordinate information, significantly improving the accuracy of positioning and response. It can not only identify and locate the device, but also link the interference module to actively cut off the operation of the hidden camera device, forming a "discovery-positioning-countermeasures" closed-loop process. By linking with the monitoring system, the device location can be intuitively marked, facilitating security personnel to quickly intervene and handle it.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for positioning and alarming a hidden camera device in combination with ambient light perception, characterized in that: include: Deploy sensor arrays, signal receiving nodes, and image acquisition devices in the target area; Obtain real-time ambient light intensity changes and ambient color distribution changes to generate ambient light intensity change trend data and ambient color distribution change trend data; The ambient light intensity change trend data and the ambient color distribution change trend data are statistically analyzed to identify abnormal ambient light changes, and the coordinate position of the abnormal ambient light change and the characteristic description of the abnormal ambient light change are output; Acquire an image at the coordinate position of the abnormal ambient light change, preprocess the image to obtain a preprocessed image, and use an image processing algorithm to extract features from the preprocessed image to obtain features of the hidden camera device; The rule base verification method is used to verify the validity of the hidden camera device features, and the candidate location and feature description of the hidden camera device are obtained; Use triangulation method to locate the candidate location of the hidden camera device and obtain the precise location coordinates of the hidden camera device; If the characteristic description of abnormal ambient light changes and the characteristic description of the hidden camera device are present at the precise location coordinates of the hidden camera device, the hidden camera device is determined to be real, and the precise location of the hidden camera device is output to the alarm system, which issues an alarm and takes interference measures; The method uses a statistical analysis method to identify abnormal ambient light changes using the ambient light intensity change trend data and the ambient color distribution change trend data, and outputs the coordinate position of the abnormal ambient light change and the abnormal ambient light change feature description. Specifically, the collected ambient light intensity change trend data and the ambient color distribution change trend data are subjected to time series modeling and multidimensional statistical analysis. First, a light intensity change curve and a corresponding color distribution histogram in frames are constructed, and the local mean, variance, and skewness statistical features are extracted using a sliding window technique. Subsequently, the threshold discrimination method and principal component analysis method were introduced to identify data points in local areas that deviated significantly from the overall trend, and potential abnormal areas were preliminarily screened out. The specific location of the abnormal areas was further determined by combining spatial coordinate mapping technology, and the characteristic information of the average brightness, hue change rate, and saturation mutation degree of the abnormal areas were extracted to construct a multi-dimensional feature vector of abnormal ambient light changes, namely the coordinate position of the abnormal ambient light changes and the characteristic description of the abnormal ambient light changes.

2. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 1, characterized in that: The abnormal ambient light change characteristics include, but are not limited to, abnormal local light intensity, abnormal light of a specific wavelength, and sudden changes in ambient color temperature.

3. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 2, characterized in that: The method acquires an image at the coordinate position of the abnormal ambient light change, preprocesses the image to obtain a preprocessed image, and uses an image processing algorithm to extract features from the preprocessed image to obtain the hidden camera device features. Specifically, the acquired image is preprocessed, including image enhancement, denoising, contrast enhancement, edge sharpening, and region of interest extraction, to obtain a preprocessed image. Subsequently, the preprocessed image is subjected to a multi-scale edge detection algorithm combined with Harris corner detection to extract the edge features and structural contours of the potential hidden camera device. Furthermore, the image is subjected to reflected light spot detection and circular / point-shaped highlight area extraction algorithms to identify lens reflections, highlight points, or geometric features that appear under abnormal lighting conditions. Combined with morphological analysis, regions with typical "lens hole" features are further screened. Finally, the hidden camera device features are obtained by constructing a feature vector and introducing a trained target detection model for classification and judgment.

4. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 3, characterized in that: The area of the "lens hole" feature includes but is not limited to a circular structure, a strong edge regularity, and a sudden change in central reflectivity; The feature vector includes structural texture, spot size, brightness gradient, shape descriptor, and spectral reflectance characteristics.

5. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 4, characterized in that: The rule base verification method is used to verify the validity of the hidden camera device features to obtain the candidate location of the hidden camera device and the hidden camera device feature description. The specific content is: the hidden camera device features are normalized, and then the similarity is calculated with the corresponding device template features in the pre-built hidden camera device feature rule base, and multiple judgment rules are set. If the hidden camera device features have matching results under multiple rules, the candidate location of the hidden camera device and the hidden camera device feature description are output.

6. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 5, characterized in that: The triangulation positioning method is used for the candidate position of the hidden camera device to obtain the precise position coordinates of the hidden camera device. Specifically, the signal receiving node located at the candidate position of the hidden camera device receives the electromagnetic signal strength in the direction of the hidden camera device, and converts the signal strength into a corresponding distance estimation value according to the RSSI and distance mapping model. A plurality of circular intersection areas are constructed with the signal receiving node as the center and the distance estimation value as the radius. The optimal overlapping point of the cross-coverage area of the plurality of nodes is calculated by the minimum mean square error method, which is the preliminary positioning point of the hidden camera device. The preliminary positioning point of the hidden camera device is combined with the vertical distribution data of the receiving antenna to further obtain the three-dimensional coordinates of the hidden camera device in space, which is the precise position coordinates of the hidden camera device.

7. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 6, characterized in that: The signal receiving node located at the candidate location of the hidden camera device receives the electromagnetic signal strength in the direction of the hidden camera device and converts the signal strength into a corresponding distance estimation value according to the RSSI and distance mapping model, which is expressed as: ； in, is the estimated distance, To calibrate the signal strength value on site, The electromagnetic signal strength in the direction of the hidden camera device. is the path loss factor.

8. The method for positioning and alarming a hidden camera device in combination with ambient light perception according to claim 7, characterized in that: The method constructs multiple circular intersection areas with the signal receiving node as the center and the distance estimation value as the radius, and calculates the optimal overlapping point of the cross-coverage area of multiple nodes by the minimum mean square error method, which is the preliminary positioning point of the hidden camera device. It is expressed as: ； in, This is the initial positioning point for the hidden camera device. is the total number of signal receiving nodes, is the distance between the i-th receiving node and the target device estimated by the RSSI mapping model, is the coordinate of the i-th signal receiving node.

9. The method for positioning and alarming a hidden camera device in combination with ambient light sensing according to claim 8, characterized in that: If the abnormal ambient light change feature description and the hidden camera device feature description exist at the precise location coordinates of the hidden camera device at the same time, the specific content of determining that the hidden camera device is real is: comparing and analyzing the ambient light change feature and the hidden camera device feature. If the ambient light change feature and the hidden camera device feature appear at the same coordinate position at the same time, the hidden camera device is determined to be real based on spatial consistency and feature superposition.

Images