Hand-held detection system for place space safety

Abstract

The invention relates to the technical field of security detection, and discloses a place space security handheld detection system, which comprises a static fingerprint construction module used for establishing a static fingerprint database and a white list shielding domain; the space-time sampling acquisition module is used for acquiring a continuous space-time sequence through the handheld device; the difference correction analysis module is used for correcting sampling and constructing a composite difference degree; the candidate clustering and screening module is used for clustering and screening candidate clusters under the constraint of the white list; the gradient approximation confirmation module is used for executing difference gradient approximation and peak domain kernel confirmation to obtain a stable peak point; and the short-distance checking and rechecking module is used for carrying out checking and positioning based on the thermal inertia and geomagnetic steady-state characteristics, and outputting a result in combination with multi-path rechecking. According to the invention, high-precision, stable and unique detection and positioning of the potential abnormal source in the complex place space are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of safety detection, and particularly relates to a place space safety handheld detection system. BACKGROUND

[0002] With the increasing requirements for safety protection in public places and key areas, how to quickly find potential abnormal sources and suspicious devices has become a problem to be solved. The common detection methods in the prior art mainly rely on fixed monitoring equipment or single type sensors, such as infrared detectors, geomagnetic sensors or video monitoring systems. However, such methods have the disadvantages of inflexible layout, high dependence on environment and high false positive rate. Fixed monitoring equipment is often limited by the field of view and installation position in complex spaces, and it is difficult to achieve full coverage; single sensor is also easily affected by environmental noise, device angle or human interference, resulting in unstable detection results.

[0003] In addition, the traditional method often fails to fuse multiple passive environmental parameters, making it difficult to comprehensively characterize the characteristics of abnormal targets. Relying on temperature alone is easily affected by season or air flow; relying on geomagnetic signals alone may be distorted by metal structures or electrical equipment interference; the electrostatic field and visible light reflection parameters are prone to instantaneous fluctuations. Due to the lack of multi-parameter joint determination and dynamic correction mechanism, the existing detection system is difficult to ensure accuracy and robustness in complex environments. SUMMARY

[0004] The application provides a place space safety handheld detection system, which solves the technical problems in the related art that relying on a single sensor, being easily affected by the environment, lacking multi-parameter fusion and review mechanism, resulting in unstable detection results, high false positive rate and difficulty in realizing unique positioning of abnormal sources.

[0005] The application provides a place space safety handheld detection system, which includes: A static fingerprint construction module is configured to establish a static fingerprint library, a white list and a shielding domain thereof, wherein the static fingerprint library is based on passive parameters collected in a preset grid under a suspected state; A space-time sampling acquisition module is configured to acquire a continuous space-time sampling sequence through a handheld device, obtain a continuous position sequence according to a handheld inspection path, and synchronously collect passive parameters at each position to form a space-time sampling sequence; A difference correction analysis module is configured to correct readings of the space-time sampling sequence and construct a composite difference degree, correct based on a posture, a walking speed and a change in a nearby environment, compare the corrected passive parameters with a static fingerprint library benchmark, and obtain a composite difference degree; The candidate cluster screening module is configured to perform candidate cluster and persistence screening under whitelist constraints, perform connectivity clustering on positions with composite difference degrees exceeding a preset threshold outside the shielding domain, perform candidate cluster centroid, maximum difference point and principal axis direction outputting according to composite difference degree drop and area condition screening, and output the candidate cluster centroid, maximum difference point and principal axis direction. The gradient approximation confirmation module is configured to perform difference gradient guided walking approximation and peak domain kernel confirmation using a handheld device, advance according to the spatial rising direction of the composite difference degree, and perform around mapping to confirm the peak domain kernel in the candidate cluster to obtain stable peak point information. The near-distance kernel verification and review module is configured to perform near-distance passive kernel verification and unique positioning based on thermal inertia characteristics and geomagnetic steady-state fluctuation characteristics, obtain a positioning result through gradient intersection of the stable peak point, and call other modules along different paths for review.

[0006] Further, the whitelist records the positions of common sources and the mean values of passive parameters. The shielding domain is obtained by merging the influence ranges of the whitelist objects. The passive parameters include temperature, geomagnetic amplitude, electrostatic potential, and visible light reflection value.

[0007] Further, a static fingerprint library and a whitelist and its shielding domain are established, wherein the static fingerprint library is based on passive parameters collected according to a preset grid under a suspicion-free state, and includes: Step 11: Clean the mobile electronic devices in the place to reach a suspicion-free state, and divide the place space according to a preset grid. Step 12: Sample each grid multiple times to collect passive parameters, and calculate the mean and standard deviation of the temperature, geomagnetic amplitude, electrostatic potential, and visible light reflection value, respectively, to take the mean and standard deviation as the reference statistical values of the grid point and store them in the static fingerprint library. Step 13: Record the positions of common sources related to passive parameters and the corresponding reference statistical values to form a whitelist, and generate a corresponding shielding area based on the influence range of each common source, and merge the shielding areas to obtain a shielding domain.

[0008] Further, a continuous spatiotemporal sampling sequence is obtained through a handheld device, a continuous position sequence is obtained according to a handheld inspection path, and passive parameters are synchronously collected at each position to form a spatiotemporal sampling sequence, including: Step 21: Walk on a preset inspection path through a handheld device, the preset inspection path being an S-shaped path covering the boundaries of the place and extending to the interior; based on inertial odometry and visual loop algorithm cooperative constraint drift error, a continuous position sequence is obtained. Step 22, at each position point of the continuous position sequence, collect in a preset time window, obtain temperature, geomagnetic amplitude, electrostatic potential and visible light reflection value in each time window respectively, and calculate window mean value to form parameter sequence; Step 23, bind the continuous position sequence with the parameter sequence one by one, and store the corresponding passive parameters with the coordinates of the position points as indexes to form a space-time sampling sequence.

[0009] Further, read the space-time sampling sequence and construct the composite difference degree, correct based on the attitude, walking speed and near neighbor environment change, compare the corrected passive parameters with the static fingerprint library benchmark to obtain the composite difference degree, including: Step 31, obtain the attitude angle of the handheld device, and compensate and correct the visible light reflection value in the space-time sampling sequence, wherein the compensation and correction uses the cosine value of the attitude angle as the correction coefficient; Step 32, determine the weight according to the walking speed at each position point, wherein the weight is obtained by calculating the ratio of the walking speed at the current position to the preset benchmark speed; in the passive parameter correction process at each position point, difference correction is performed in combination with the average value of the sampling points within the preset sampling radius range centered on the position point; Step 33, compare the corrected passive parameters with the corresponding benchmark values of the static fingerprint library, wherein the comparison uses the standard deviation of each passive parameter in the static fingerprint library as a normalization factor, and calculates the square root of the sum of the squares of the difference between each passive parameter and the corresponding benchmark value to obtain the composite difference degree.

[0010] Further, the candidate cluster screening module includes: Step 41, screen the position points outside the shielding domain whose composite difference degree exceeds a preset difference degree threshold to obtain a candidate point set, wherein the preset threshold is the sum of the mean value of the composite difference degree and the standard deviation of the preset multiple; Step 42, cluster the candidate point set according to spatial connectivity to obtain a candidate cluster set, wherein the spatial connectivity is determined based on whether the Euclidean distance between the candidate points is less than a preset neighborhood radius; and calculate the centroid, maximum difference point and principal axis direction of each candidate cluster, wherein the principal axis direction is obtained by performing eigenvalue decomposition on the covariance matrix of the candidate cluster point set; Step 43, calculate the persistence of the candidate cluster, and screen in combination with the area condition to output the centroid, maximum difference point and principal axis direction of the candidate cluster that meets the condition; the persistence is the difference between the maximum value of the composite difference degree in the candidate cluster and the minimum value of the composite difference degree at the convex hull boundary of the candidate cluster, and the area condition is obtained by the number of points in the candidate cluster being greater than or equal to a preset threshold.

[0011] Further, the difference gradient guided walking approximation and peak domain kernel confirmation include: Step 51, calculating the ascending direction based on the gradient of the composite difference degree in the candidate cluster, selecting the position direction with the fastest ascending as the moving direction, updating the position points and forming the approaching path sequence by carrying the handheld device to walk step by step along the moving direction, the ascending direction is obtained by differentiating the composite difference degree of adjacent position points in the candidate cluster, and the step-by-step walking adopts an adaptive step length adjusted according to the gradient amplitude of the composite difference degree; Step 52, when the approaching path approaches the local peak region of the candidate cluster, sampling according to a closed path around the peak, forming a surrounding sampling set, and drawing contour lines after interpolating the composite difference degree of the surrounding sampling set on the plane, determining the boundary of the peak domain core according to the contour line shape; the closed path is a ring path with the local peak point as the center; Step 53, determining the point with the maximum composite difference degree in the peak domain core as the candidate peak point, and repeatedly executing sampling under different surrounding mapping paths, if the candidate peak point is detected under at least two surrounding mapping paths, and the spatial position difference is less than a preset consistency threshold, the candidate peak point is determined as a stable peak point, and the stable peak point information is output.

[0012] Further, in step 52, the determination of the boundary of the peak domain core is based on that the change rate of the area of the contour closed region is less than a preset change rate threshold.

[0013] Further, the execution of the near-distance passive verification includes: Step 61, collecting continuous temperature data at the stable peak point to form a temperature sampling sequence, and calculating the temperature change rate with time by differentiating adjacent sampling points, when the temperature change rate is lower than a preset thermal inertia threshold, it is determined that the peak point has thermal inertia characteristics; Step 62, collecting continuous geomagnetic amplitude data at the stable peak point to form a geomagnetic sampling sequence, and calculating the mean and standard deviation of the geomagnetic amplitude based on a sliding time window, when the standard deviation is less than a preset geomagnetic steady-state threshold, it is determined that the peak point has geomagnetic steady-state fluctuation characteristics; Step 63, jointly verifying the thermal inertia characteristic determination result and the geomagnetic steady-state fluctuation determination result, confirming the peak point as an effective peak point only when both satisfy, otherwise determining it as an invalid peak point; when there are multiple similar effective peak points, taking the spatial Euclidean distance less than a preset distance threshold as the similarity determination condition, and selecting the peak point with the maximum composite difference degree as the only peak point, and outputting the unique effective peak point information.

[0014] Further, the gradient intersection of the stable peak point is obtained to obtain the positioning result, and other modules are called along different paths for rechecking, including: Step 71, based on the unique effective peak point information and the continuous position sequence of the handheld device, coordinate reference unification and error modeling are performed, the unique effective peak point is converted into a candidate positioning coordinate in a global coordinate system, and an error radius is obtained by calculating the square root of the sampling variance of the peak point; Step 72, a verification observation point is uniformly distributed in the neighborhood of the candidate positioning coordinate, a short-range walk is performed in the maximum rising direction of the composite difference degree at each verification observation point, and a plurality of verification direction lines are formed; the intersection point is obtained as the positioning result by performing least square calculation on the verification direction lines; Step 73, the spatio-temporal sampling acquisition module, the difference correction analysis module, the candidate clustering screening module and the gradient approximation confirmation module are repeatedly executed along at least two different inspection paths from the original inspection path, step 72 is re-executed in the neighborhood of the candidate positioning coordinate, if the Euclidean distance between the positioning results obtained multiple times is less than a preset verification threshold, the final positioning result is confirmed, otherwise it is marked as needing to be rechecked; the different inspection paths are mutually orthogonal S-shaped paths.

[0015] The beneficial effects of the present application are that: the present application effectively avoids the false and missed report problems caused by single sensor dependence through joint collection and composite difference degree calculation of multiple passive parameters such as temperature, geomagnetic amplitude, electrostatic potential and visible light reflection value; the accuracy and stability of the sampling data are significantly improved by combining attitude angle compensation, speed weight and neighborhood difference correction mechanism; common interference sources can be automatically identified and removed by introducing the constraints of the white list and the shielding domain, reducing the influence of environmental noise on the detection result; the uniqueness and reliability of the abnormal source positioning result are ensured by using multi-level processing methods such as candidate clustering, difference gradient guidance, peak domain kernel confirmation and stable peak point screening; the consistency of the positioning result is further verified by the multi-path verification mechanism, and the robustness of the system under different places and inspection conditions is improved. Overall, the present application can realize high-precision, high-stability and strong-adaptability place space safety detection. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a module schematic diagram of a place space safety handheld detection system of the present application. DETAILED DESCRIPTION

[0017] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can be changed in function and arrangement without departing from the scope of the present description. Various processes or components can be omitted, substituted, or added according to desired implementations. Additionally, features described with respect to some examples can be combined in other examples.

[0018] As shown in Figure 1 A place space safety handheld detection system comprises: A static fingerprint construction module is configured to establish a static fingerprint library and a whitelist and a shielding domain, wherein the static fingerprint library is based on passive parameters collected in a preset grid under a non-suspicious state; A space-time sampling acquisition module is configured to acquire a continuous space-time sampling sequence through a handheld device, obtain a continuous position sequence according to a handheld inspection path, and synchronously acquire passive parameters at each position to form a space-time sampling sequence; A difference correction analysis module is configured to correct readings and construct a composite difference degree for the space-time sampling sequence, correct based on a posture, a walking speed, and a change in a nearby environment, compare the corrected passive parameters with a static fingerprint library benchmark, and obtain a composite difference degree; A candidate cluster screening module is configured to perform candidate cluster and persistence screening under whitelist constraints, perform connectivity clustering on positions with a composite difference degree exceeding a preset threshold outside the shielding domain, and screen according to a difference drop and an area condition to output a centroid, a maximum difference point, and a principal axis direction of a candidate cluster; A gradient approximation confirmation module is configured to perform a walking approximation guided by a difference gradient and a peak domain kernel confirmation using a handheld device, advance according to a spatial rising direction of the composite difference degree, and perform a surrounding mapping confirmation of a peak domain kernel within a candidate cluster to obtain stable peak point information; A near-distance kernel verification and rechecking module is configured to perform a near-distance passive verification and a unique positioning based on thermal inertia characteristics and geomagnetic steady-state fluctuation characteristics, obtain a positioning result through a gradient intersection of stable peak points, and call other modules along different paths for rechecking.

[0019] In an embodiment of the present application, the whitelist records a position and a passive parameter mean value of a common source; the common source refers to a passive physical interference source that exists long-term and is stable and unchanged in an inspection place, such as a fixedly installed lighting fixture, a metal support, a ventilation opening, or a large device, which produces a stable background response in the passive parameter acquisition process. By recording the position coordinates and the corresponding passive parameter mean value of such a common source in the whitelist, the common source can be identified and shielded in a subsequent detection process, thereby avoiding being misjudged as an abnormal target.

[0020] The shielding domain is obtained by merging an influence range of a whitelist object; the shielding domain is used to eliminate differences caused by common sources during abnormal detection, thereby improving the accuracy and stability of detection.

[0021] The passive parameters include temperature, geomagnetic amplitude, electrostatic potential, and visible light reflection value; the visible light reflection value refers to a reflection intensity value of a target position to incident light in a visible light band.

[0022] In an embodiment of the present application, a static fingerprint library and a whitelist and shielding area thereof are established, wherein the static fingerprint library is based on passive parameters collected in a preset grid under a suspicious-free state, including: Step 11, clean up the mobile electronic devices in the place to reach a suspicious-free state, and divide the space of the place according to a preset grid; wherein the suspicious-free state means that there is no active abnormal electronic device or suspicious interference source in the place, only fixed structures and stable background signals are reserved; Step 12, sample each grid multiple times to collect passive parameters, and calculate the mean and standard deviation of the temperature, geomagnetic amplitude, electrostatic potential and visible light reflection value respectively, and take the mean and standard deviation as the reference statistical value of the grid point and store it in the static fingerprint library; the static fingerprint library is used as a reference for the subsequent detection process to calculate the composite difference degree; Step 13, record the positions of common sources related to passive parameters and the corresponding reference statistical values to form a whitelist, and generate corresponding shielding areas based on the influence range of each common source, and combine the shielding areas to obtain a shielding area; by setting the whitelist and the shielding area, the interference of the common source on the detection result can be effectively avoided, thereby improving the accuracy and robustness of the detection.

[0023] In an embodiment of the present application, a continuous spatio-temporal sampling sequence is obtained by a handheld device, a continuous position sequence is obtained according to a handheld inspection path, and passive parameters are synchronously collected at each position to form a spatio-temporal sampling sequence, including: Step 21, walk along a preset inspection path by a handheld device, the preset inspection path is an S-shaped path covering the boundary of the place and extending to the interior; based on the cooperative constraint of inertial odometry and visual loop algorithm to drift error, a continuous position sequence is obtained; specifically, an operator carries a handheld device to walk along a preset inspection path in the place, the inspection path is designed as an S-shaped path covering the boundary of the place and extending to the interior region, to ensure the integrity of the space coverage. In the inspection process, the handheld device simultaneously relies on inertial odometry and visual loop algorithm to calculate the position, wherein the inertial odometry is used to calculate the relative displacement according to the accelerometer and gyroscope signals, and the visual loop algorithm identifies repeated scenes through the environment images captured by the camera, thereby constraining and correcting the cumulative drift of the inertial odometry. Through the cooperative action of the two, a continuous position sequence with controlled error can be obtained.

[0024] Step 22, at each position point of the continuous position sequence, collect in a preset time window, obtain the temperature, geomagnetic amplitude, electrostatic potential and visible light reflection value in each time window respectively, and calculate the window mean value to form a parameter sequence; by calculating the window mean value, the interference caused by instantaneous fluctuation is reduced.

[0025] Step 23, binding the continuous position sequence with the parameter sequence one by one, and storing the corresponding passive parameters with the coordinates of the position points as indexes to form a time-space sampling sequence. The time-space sampling sequence established in this way can accurately reflect the environmental passive characteristics of each position point on the inspection path, eliminate instantaneous noise in the time dimension, and keep the accurate correspondence between the coordinates and the parameters in the space dimension.

[0026] In an embodiment of the present application, the time-space sampling sequence is readjusted and a composite difference degree is constructed. The readjustment is based on the posture, walking speed and change of the adjacent environment. The readjusted passive parameters are compared with the reference of the static fingerprint library to obtain the composite difference degree, which includes: Step 31, the posture angle of the handheld device is obtained, and the visible light reflection value in the time-space sampling sequence is compensated and adjusted. The compensation and adjustment adopts the cosine value of the posture angle as the correction coefficient. The posture angle refers to the inclination angle of the device relative to the horizontal plane, which is measured by the accelerometer. Since the collection results of the visible light reflection of the device at different angles will be affected, it is necessary to compensate and adjust the visible light reflection value in the time-space sampling sequence. In this embodiment, the cosine value of the posture angle is used as the correction coefficient, which is multiplied by the original visible light reflection value, so as to offset the measurement deviation caused by the inclination of the device, and realize the posture compensation of the visible light reflection data. Step 32, the weight is determined according to the walking speed at each position point. The weight is obtained by calculating the ratio of the walking speed at the current position to the preset reference speed. When the actual walking speed is faster or slower than the reference speed, the weight will be adjusted accordingly, so as to reflect the stability of the sampling data in the time dimension. In the passive parameter adjustment process at each position point, the average value of the sampling points within the preset sampling radius centered on the position point is combined for differential adjustment, so as to correct the random interference caused by the local environmental fluctuation. Step 33, the adjusted passive parameters are compared with the corresponding reference values of the static fingerprint library. The comparison adopts the standard deviation of each passive parameter in the static fingerprint library as the normalization factor, and calculates the square root of the sum of the squares of the difference between each passive parameter and the corresponding reference value to obtain the composite difference degree. The composite difference degree is a comprehensive index reflecting the difference degree between the actual sampling value and the reference environment. The larger the value is, the higher the degree of deviation of the passive characteristics of the position point from the suspicious state is.

[0027] Through the above process, the original sampling data can be adjusted in multiple dimensions in this embodiment, including the elimination of the posture influence, the compensation of the speed difference and the smoothing correction of the environment neighborhood, so as to ensure the objectivity and consistency of the comparison.

[0028] In an embodiment of the present application, the candidate cluster screening module includes: Step 41, screening the position points with composite difference exceeding a preset difference threshold outside the shielding domain to obtain a candidate point set, the preset threshold being the sum of the mean value of the composite difference and the standard deviation multiplied by a preset factor; this statistical method can dynamically adapt to environmental fluctuations and reduce over-sensitivity or omissions.

[0029] Step 42, clustering the candidate point set according to spatial connectivity to obtain a candidate cluster set, the spatial connectivity being determined based on whether the Euclidean distance between candidate points is less than a preset neighborhood radius; and calculating the centroid, maximum difference point and principal axis direction of each candidate cluster, the principal axis direction being obtained by eigenvalue decomposition of the covariance matrix of the candidate cluster point set; specifically, the spatial connectivity refers to whether the candidate points are close enough in space, and the determination method is based on the Euclidean distance, that is, by calculating the straight-line distance between two candidate points in two-dimensional or three-dimensional space coordinates, when the distance is less than the preset neighborhood radius, it is considered that the two are connected. The centroid refers to the arithmetic mean of the coordinates of all points in the cluster, which is used to reflect the position center of the cluster; the maximum difference point refers to the point with the highest composite difference value in the cluster, which is used to indicate the most significant position of the anomaly.

[0030] Step 43, calculating the persistence of the candidate cluster and screening in combination with the area condition to output the centroid, maximum difference point and principal axis direction of the candidate cluster that meets the condition; the persistence being the difference between the maximum composite difference value in the candidate cluster and the minimum composite difference value at the convex hull boundary of the candidate cluster, wherein the convex hull refers to the smallest convex polygon or convex polyhedron that can cover all points in the cluster. In this way, the difference amplitude between the abnormal signal in the cluster and the boundary noise can be measured, and the greater the persistence, the more stable the anomaly; the area condition is obtained by the number of points in the candidate cluster being greater than or equal to a preset threshold.

[0031] Through the above steps, the candidate clustering screening module can effectively eliminate environmental noise and isolated interference in a large number of sampling points, and retain regions with significant spatial continuity and stable difference, thereby improving the accuracy and reliability of subsequent anomaly positioning.

[0032] In an embodiment of the present application, the difference gradient guided walking approximation and peak domain kernel confirmation include: Step 51, calculating an ascending direction based on a gradient of the composite difference degree in the candidate cluster, selecting a position direction with the fastest ascending as a moving direction, updating the position point and forming a sequence of approaching paths by walking along the moving direction with the handheld device, the ascending direction being obtained by differentiating the composite difference degree of adjacent position points in the candidate cluster, the walking being with an adaptive step length adjusted according to the gradient amplitude of the composite difference degree, i.e. automatically adjusting the step length according to the gradient amplitude of the composite difference degree, using a smaller step length to improve the precision when the gradient is larger and using a larger step length to improve the efficiency when the gradient is smaller, in this way, the local peak region can be approached step by step along the most significant ascending direction. The gradient of the composite difference degree refers to the direction of change of the composite difference degree in the spatial distribution, which is obtained by differentiating the composite difference degree of adjacent position points in the candidate cluster.

[0033] Step 52, when the approaching path approaches the local peak region of the candidate cluster, sampling around the peak according to a closed path to form a surrounding sampling set, and drawing contour lines after interpolating the composite difference degree of the surrounding sampling set on the plane to determine the boundary of the peak domain nucleus according to the contour line shape; the closed path is an annular path with the local peak point as the center, so that the sampling points are uniformly distributed around the peak; the interpolation processing uses linear interpolation to estimate continuous values, thereby generating continuous contour lines; the contour line shape reflects the distribution characteristics of the difference degree in space. The determination of the boundary of the peak domain nucleus is based on the change rate of the area of the closed region being less than a preset change rate threshold.

[0034] Step 53, determining the point with the maximum composite difference degree in the peak domain nucleus as a candidate peak point, and repeatedly performing sampling under different surrounding mapping paths, if the candidate peak point is detected under at least two surrounding mapping paths and the spatial position difference is less than a preset consistency threshold, the candidate peak point is determined as a stable peak point, and the stable peak point information is output. The stable peak point information can include the spatial position coordinates of the peak point, the corresponding composite difference degree value, and the determination results of the thermal inertia feature and the geomagnetic steady-state feature; the preset consistency threshold refers to the maximum spatial deviation allowed for the peak point position under different mapping paths.

[0035] Through the above design, the embodiment can effectively exclude environmental noise interference in a complex place, realize rapid approximation guided by the gradient, and improve the stability and accuracy of peak point positioning by combining surrounding mapping and repeated verification, thereby providing reliable initial candidate points for subsequent near-passive verification and unique positioning.

[0036] In an embodiment of the present application, performing near-passive verification includes: Step 61, collect continuous temperature data at the stable peak point to form a temperature sampling sequence, and calculate the temperature change rate with time by difference calculation between adjacent sampling points, and determine that the peak point has thermal inertia characteristics when the temperature change rate is lower than a preset thermal inertia threshold, which is determined by baseline temperature experiments in a non-suspicious state; Step 62, collect continuous geomagnetic amplitude data at the stable peak point to form a geomagnetic sampling sequence, and calculate the mean and standard deviation of the geomagnetic amplitude based on a sliding time window, and determine that the peak point has geomagnetic steady-state fluctuation characteristics when the standard deviation is less than a preset geomagnetic steady-state threshold, which is a preset multiple of the standard deviation of the geomagnetic amplitude in the static fingerprint library, so that the steady-state determination range can be adaptively adjusted under different environmental conditions; Step 63, jointly verify the thermal inertia characteristic determination result and the geomagnetic steady-state fluctuation determination result, and only when both meet the requirements, confirm the peak point as a valid peak point, otherwise determine it as an invalid peak point; when there are multiple similar valid peak points, take the spatial Euclidean distance less than a preset distance threshold as the similarity determination condition, and select the peak point with the largest composite difference degree from them as the only peak point, and output the unique valid peak point information.

[0037] Through the above steps, the embodiment can use thermal inertia characteristics and geomagnetic steady-state fluctuation characteristics to double-check the peak points, thereby significantly improving the reliability of abnormal source identification. Joint verification avoids misjudgment caused by a single feature, and the uniqueness processing of similar peak points ensures the uniqueness and accuracy of the final positioning result.

[0038] In an embodiment of the present application, the positioning result is obtained by gradient intersection of stable peak points, and other modules are called along different paths for rechecking, including: Step 71, based on the unique valid peak point information and the continuous position sequence of the handheld device, perform coordinate reference unification and error modeling, convert the unique valid peak point into a candidate positioning coordinate in the global coordinate system, and obtain an error radius by calculating the square root of the sampling variance of the peak point, which is used to quantify the uncertainty range of the peak point positioning; wherein the coordinate reference unification refers to converting the peak point position in the relative coordinate system to the global coordinate system, so as to ensure that the processing results of different paths and different modules have consistent spatial reference; Step 72, set up calibration observation points uniformly in the neighborhood of the candidate positioning coordinate, and perform short-range walking along the maximum rising direction of the composite difference degree at each calibration observation point to form multiple calibration direction lines; obtain the intersection point as the positioning result by least square calculation of the calibration direction lines, that is, by minimizing the sum of squares of distances between the calibration direction lines and the intersection point; the maximum rising direction of the composite difference degree is the direction with the most significant increase in the gradient of the composite difference degree around the point, which is used to indicate the spatial expansion trend of the abnormal source.

[0039] Step 73, repeat the execution of the space-time sampling acquisition module, the difference correction analysis module, the candidate cluster screening module and the gradient approximation confirmation module along at least two different inspection paths from the original inspection path, re-execute step 72 in the neighborhood of the candidate positioning coordinates, if the Euclidean distance between the positioning results obtained multiple times is less than the preset review threshold, the final positioning result is confirmed, otherwise it is marked as needing review; the different inspection paths are mutually orthogonal S-shaped paths to ensure multi-angle coverage and cross-validation of the space region.

[0040] Through the above design, the embodiment can effectively improve the uniqueness and reliability of positioning on the basis of preliminary positioning, combined with the review mechanism of multiple paths, multiple modules and multiple observation points. The unified coordinate reference and error modeling provide quantitative constraints for the results, and the checking observation points and least squares calculation enhance the robustness of geometric intersection, while the multi-path review mechanism ensures that the results are not dependent on the accidental characteristics of a single path, thereby overall improving the adaptability and credibility of the detection system in complex site environments.

[0041] It should be noted that the setting of the interval and the threshold size is for easy comparison, and the size of the threshold depends on the number of sample data and the base number set by the person skilled in the art for each group of sample data, as long as it does not affect the proportional relationship between the parameters and the quantized values. And the above formula is a calculation of the dimensionless value, and the formula is obtained by software simulation of a large amount of data to obtain a formula closest to the real situation, and the preset parameters in the formula are set by the person skilled in the art according to the actual situation.

[0042] The embodiments of the present application are described above, but the present application is not limited to the above specific embodiments, and the above specific embodiments are only illustrative and not limiting, and a person of ordinary skill in the art can make many forms under the inspiration of the present embodiment, which are all within the protection of the present embodiment.

Claims

1. A venue space safety handheld detection system, characterized by, include: A static fingerprint building module is used to establish a static fingerprint library and a whitelist and its shielding domain. The static fingerprint library collects passive parameters according to a preset grid in a non-suspicious state. The spatiotemporal sampling acquisition module is used to obtain a continuous spatiotemporal sampling sequence through a handheld device, obtain a continuous position sequence according to the handheld inspection path, and synchronously collect passive parameters at each position to form a spatiotemporal sampling sequence; The difference correction analysis module is used to correct the readings of the spatiotemporal sampling sequence and construct a composite difference. The correction is based on the changes in posture, walking speed and the surrounding environment. The corrected passive parameters are compared with the static fingerprint library benchmark to obtain the composite difference. The candidate cluster screening module is used to screen candidate clusters and persistence under the whitelist constraint. It performs connectivity clustering on locations outside the shielded domain where the composite difference exceeds the preset threshold. It screens based on the composite difference drop and area conditions and outputs the centroid, maximum difference point, and main axis direction of the candidate cluster. The gradient approximation confirmation module is used to perform walking approximation and peak kernel confirmation guided by the difference gradient using a handheld device. It advances according to the spatial ascending direction of the composite difference degree and performs surround mapping within the candidate cluster to confirm the peak kernel and obtain stable peak point information. The close-range verification and recheck module is used to perform close-range passive verification and unique positioning based on thermal inertia characteristics and geomagnetic steady-state fluctuation characteristics. The positioning result is obtained through the gradient intersection of stable peak points, and other modules are called along different paths for rechecking.

2. A hand-held detection system for security of a site space according to claim 1, wherein, The whitelist records the locations of common sources and passive parameter means; The shielding domain is obtained by merging the influence ranges of the whitelist objects; The passive parameters include: temperature, geomagnetic amplitude, electrostatic potential, and visible light reflection value.

3. A hand-held detection system for security of a site space according to claim 1, wherein, Establish a static fingerprint library, whitelist, and shielding domain. The static fingerprint library collects passive parameters based on a preset grid in a non-suspicious state, including: Step 11: Clear the mobile electronic devices in the venue until they are clear of suspicion, and divide the venue space into preset grids; Step 12: Sampling each grid multiple times, collecting passive parameters, and calculating the mean and standard deviation of temperature, geomagnetic amplitude, electrostatic potential, and visible light reflectance respectively. The mean and standard deviation are used as the benchmark statistics of the grid point and stored in the static fingerprint library. Step 13: Record the locations of common sources related to passive parameters and the corresponding baseline statistical values ​​to form a whitelist, generate corresponding shielding areas based on the impact range of each common source, and merge the shielding areas to obtain a shielding domain.

4. A hand-held detection system for security of a site space according to claim 1, wherein, A continuous spatiotemporal sampling sequence is obtained through a handheld device. A continuous position sequence is obtained based on the handheld inspection path. Passive parameters are synchronously collected at each position to form a spatiotemporal sampling sequence, including: Step 21: Use a handheld device to walk along a preset inspection path, which is an S-shaped path that covers the site boundary and extends into the interior; based on the inertial odometry and visual loop closure algorithm, the drift error is constrained to obtain a continuous position sequence; Step 22, at each position point of the continuous position sequence, collect in a preset time window, obtain temperature, geomagnetic amplitude, electrostatic potential and visible light reflection value in each time window respectively, and calculate window mean value to form parameter sequence; Step 23, bind the continuous position sequence with the parameter sequence one by one, and store the corresponding passive parameters with the coordinates of the position points as indexes to form a space-time sampling sequence.

5. A hand-held detection system for security of a site space according to claim 1, wherein, Read the space-time sampling sequence and construct the composite difference degree, correct based on attitude, walking speed and near neighbor environment change, compare the corrected passive parameters with the static fingerprint library benchmark to obtain the composite difference degree, including: Step 31, obtain the attitude angle of the handheld device, and compensate and correct the visible light reflection value in the space-time sampling sequence, wherein the compensation and correction uses the cosine value of the attitude angle as the correction coefficient; Step 32, determine the weight according to the walking speed at each position point, wherein the weight is obtained by calculating the ratio of the walking speed at the current position to the preset benchmark speed; during the passive parameter correction at each position point, difference correction is performed in combination with the average value of the sampling points within the preset sampling radius range centered on the position point; Step 33, compare the corrected passive parameters with the corresponding benchmark values of the static fingerprint library, wherein the comparison uses the standard deviation of each passive parameter in the static fingerprint library as the normalization factor, and calculates the square root of the sum of the squares of the difference between each passive parameter and the corresponding benchmark value to obtain the composite difference degree.

6. A hand-held detection system for security of a site space according to claim 1, wherein, The candidate cluster screening module includes: Step 41, screen the position points with a composite difference degree exceeding a preset difference degree threshold outside the shielding domain to obtain a candidate point set, wherein the preset threshold is the sum of the mean value of the composite difference degree and the standard deviation multiplied by a preset multiple; Step 42, cluster the candidate point set according to spatial connectivity to obtain a candidate cluster set, wherein the spatial connectivity is determined based on whether the Euclidean distance between candidate points is less than a preset neighborhood radius; and calculate the centroid, maximum difference point and principal axis direction of each candidate cluster; the principal axis direction is obtained by performing eigenvalue decomposition on the covariance matrix of the candidate cluster point set; Step 43, calculate the persistence of the candidate cluster, and screen in combination with the area condition to output the centroid, maximum difference point and principal axis direction of the candidate cluster that meets the condition; the persistence is the difference between the maximum composite difference degree in the candidate cluster and the minimum composite difference degree at the convex hull boundary of the candidate cluster, and the area condition is obtained by the number of points in the candidate cluster being greater than or equal to a preset threshold.

7. A hand-held detection system for security of a site space according to claim 1, wherein, The difference gradient guided walking approximation and peak domain kernel confirmation includes: Step 51, calculate the rising direction based on the gradient of the composite difference degree in the candidate cluster, select the position direction with the fastest rising as the walking direction, and gradually walk along the walking direction by carrying the handheld device to update the position points and form an approximation path sequence; the rising direction is obtained by performing difference calculation on the composite difference degrees of adjacent position points in the candidate cluster, and the gradual walking uses an adaptive step length adjusted according to the gradient amplitude of the composite difference degree; Step 52, when the approaching path approaches the local peak area of the candidate cluster, sampling is performed around the peak according to a closed path to form a surrounding sampling set, and the composite difference degree of the surrounding sampling set is interpolated on the plane to draw contour lines, and the boundary of the peak domain core is determined according to the contour line shape; the closed path is a ring path with the local peak point as the center; Step 53, the point with the maximum composite difference degree in the peak domain core is determined as a candidate peak point, and the sampling is repeatedly performed under different surrounding mapping paths, if the candidate peak point is detected under at least two surrounding mapping paths, and the spatial position difference is less than a preset consistency threshold, the candidate peak point is determined as a stable peak point, and the stable peak point information is output.

8. A hand-held detection system for security of a site space according to claim 7, wherein, In step 52, the determination of the boundary of the peak domain core is based on that the change rate of the area of the contour closed region is less than a preset change rate threshold.

9. A hand-held detection system for security of a site space according to claim 1, wherein, The execution of the near-distance passive verification includes: Step 61, continuous temperature data is collected at the stable peak point to form a temperature sampling sequence, and the temperature change rate with time is calculated by difference calculation of adjacent sampling points, and when the temperature change rate is lower than a preset thermal inertia threshold, it is determined that the peak point has thermal inertia characteristics; Step 62, continuous geomagnetic amplitude data is collected at the stable peak point to form a geomagnetic sampling sequence, and the mean value and standard deviation of the geomagnetic amplitude are calculated based on a sliding time window, and when the standard deviation is less than a preset geomagnetic steady-state threshold, it is determined that the peak point has geomagnetic steady-state fluctuation characteristics; Step 63, the thermal inertia characteristic determination result and the geomagnetic steady-state fluctuation determination result are jointly verified, and only when both satisfy the condition, the peak point is confirmed as a valid peak point, otherwise it is determined as an invalid peak point; when there are multiple similar valid peak points, the spatial Euclidean distance less than a preset distance threshold is taken as a similarity determination condition, and the peak point with the maximum composite difference degree is selected as the only peak point, and the unique valid peak point information is output.

10. A hand-held detection system for security of a site space according to claim 1, characterized in that, The positioning result is obtained by gradient intersection of the stable peak point, and other modules are called along different paths for rechecking, including: Step 71, based on the unique valid peak point information and the continuous position sequence of the handheld device, coordinate reference unification and error modeling are performed, the unique valid peak point is converted into a candidate positioning coordinate in a global coordinate system, and the error radius is obtained by calculating the square root of the sampling variance of the peak point; Step 72, taking the candidate positioning coordinate as the center, a plurality of calibration observation points are uniformly distributed in the neighborhood of the candidate positioning coordinate, and a short-range walk is performed in the maximum rising direction of the composite difference degree at each calibration observation point to form a plurality of calibration direction lines; the intersection point obtained by least square calculation of the calibration direction lines is taken as the positioning result; Step 73, the spatio-temporal sampling acquisition module, the difference correction analysis module, the candidate clustering screening module and the gradient approximation confirmation module are repeatedly executed along at least two inspection paths different from the original inspection path, step 72 is re-executed in the neighborhood of the candidate positioning coordinate, if the Euclidean distances between the positioning results obtained multiple times are less than a preset rechecking threshold, the final positioning result is confirmed, otherwise it is marked as needing rechecking; the different inspection paths are mutually orthogonal S-shaped paths.

Images