Wall-crossing compliance judgment method based on scene graph and trajectory analysis

Abstract

Embodiments of the present application provide a wall-crossing compliance judgment method based on scene graph and trajectory analysis. The method comprises obtaining building structure information and indoor positioning trajectory data of a region to be judged; generating an overall obstacle graph based on the building structure information; performing edge detection and feature fusion processing on the overall obstacle graph to generate boundary point coordinate indexes; performing iterative convolution operation on the boundary point coordinate indexes and the overall obstacle graph to generate an obstacle weight graph; calculating a trajectory evaluation index based on the indoor positioning trajectory data and the obstacle weight graph; and analyzing the trajectory evaluation index based on a preset compliance verification process to generate a compliance result of the indoor positioning trajectory data. In this way, the accuracy of indoor positioning trajectory analysis can be improved.

Description

Technical Field

[0001] The embodiments of this application relate to the field of indoor positioning and building information processing, and in particular to a method for judging compliance of wall crossing and boundary crossing based on scene map and trajectory analysis. Background Technology

[0002] In indoor environments, outdoor positioning technologies such as GPS are difficult to apply effectively due to signal attenuation and obstruction. Indoor positioning technology has become one of the core supporting technologies in fields such as smart shopping malls, industrial production, medical care, and robot navigation. Currently, the industry has formed several mainstream technical paths based on electromagnetic waves, inertial navigation, visual positioning, magnetic field positioning, and multi-source fusion. However, all methods generally face an irreconcilable contradiction between accuracy, complexity, and resource consumption.

[0003] Indoor positioning errors stem from multiple factors, including multipath effects during signal propagation, electromagnetic / magnetic interference in complex environments, noise in the sensor hardware itself, and mismatches between the algorithm model and the actual scene. This leads to problems such as high system complexity, difficulty in data synchronization, and weak environmental adaptability in existing fusion solutions, making it difficult to achieve a balance between accuracy, stability, and cost. To assess the compliance of indoor positioning trajectories and determine whether the generated errors are within acceptable limits, existing research and applications have proposed various judgment methods. Some technologies rely on constructing complex mathematical models, such as establishing an elliptical uncertain region based on two reference points to evaluate trajectory errors. However, the uncertain region constructed by such methods is difficult to accurately depict the complex shapes between reference points, and the size of the elliptical region is difficult to determine, resulting in reduced accuracy in judging trajectory compliance in practical applications. Other methods are based on machine learning, predicting the rationality of trajectories by learning from large amounts of historical positioning data. However, these methods not only require significant time and computational resources for data training, but also suffer from insufficient generalization ability when facing dynamically changing indoor environments, making them prone to misjudgments. Summary of the Invention

[0004] According to an embodiment of this application, a method for judging compliance of wall crossing and boundary crossing based on scene graph and trajectory analysis is provided, which can improve the accuracy of indoor positioning trajectory analysis.

[0005] In a first aspect of this application, a method for determining compliance of wall crossing and boundary intrusion based on scene graph and trajectory analysis is provided. The method includes:

[0006] Obtain building structure information and indoor positioning trajectory data for the area to be judged;

[0007] A comprehensive obstacle map is generated based on the aforementioned building structure information;

[0008] Edge detection and feature fusion are performed on the overall obstacle map to generate boundary point coordinate indices;

[0009] Perform iterative convolution operations on the boundary point coordinate indices and the overall obstacle map to generate an obstacle weight map;

[0010] The trajectory evaluation index is calculated based on the indoor positioning trajectory data and the obstacle weight map;

[0011] The trajectory evaluation indicators are analyzed based on a preset compliance verification process to generate compliance results for the indoor positioning trajectory data.

[0012] In one possible implementation, generating the overall obstacle map based on the building structure information includes:

[0013] The building walls and building boundaries are determined based on the aforementioned building structure information;

[0014] A rigid wall diagram is generated based on the building walls and the building boundaries;

[0015] Obtain a binarization allocation scheme, wherein the binarization allocation scheme sets the area outside the building walls and building boundaries as the first pixel value and the passable area as the second pixel value;

[0016] The rigid wall map is binarized based on the aforementioned binarization allocation scheme to generate an overall obstacle map.

[0017] In one possible implementation, the step of performing edge detection and feature fusion processing on the overall obstacle map to generate boundary point coordinate indices includes:

[0018] Obtain the target image processing direction;

[0019] The overall obstacle map is deconvolved based on the convolution kernels in the target image processing direction to obtain a multi-directional feature map;

[0020] The multi-directional feature maps are spatially aligned and fused to generate a processed feature map;

[0021] The processed feature map is pooled, and a Boolean matrix of marked boundary points is generated by filtering through a preset threshold.

[0022] Data extraction is performed on the Boolean matrix to obtain the coordinate indices of the boundary points.

[0023] In one possible implementation, the iterative convolution operation on the boundary point coordinate indices and the overall obstacle map to generate an obstacle weight map includes:

[0024] Based on the boundary point coordinate index, the boundary position pixels in the overall obstacle map are weighted to obtain an initial weighted map;

[0025] The initial weighted image is convolved using a preset convolution kernel to generate a convolution result.

[0026] Add 1 to the convolution result and take the reciprocal to obtain the first intermediate feature map;

[0027] The overall obstacle map is convolved to generate a neighborhood summation mapping, and the pixels at the coordinate index positions of the boundary points are normalized to obtain a second intermediate feature map.

[0028] An initial weight map is generated based on the first intermediate feature map, the second intermediate feature map, the overall obstacle map, and a preset number of iterations.

[0029] Boundary post-processing is performed on the initial weight map to generate an obstacle weight map.

[0030] In one possible implementation, the step of performing boundary post-processing on the initial weight map to generate a barrier weight map includes:

[0031] Based on the rigid wall diagram, determine the nearest neighbor region within a preset pixel length range inside and outside the boundary;

[0032] Interference filtering is performed on the neighboring regions and the corresponding positions in the initial weight map;

[0033] The weight values ​​in the preset far boundary region of the initial weight map are set to the maximum weight values ​​in the initial weight map to obtain the final obstacle weight map.

[0034] In one possible implementation, the calculation of trajectory evaluation metrics based on the indoor positioning trajectory data and the obstacle weight map includes:

[0035] Obtain outbound limit information and indicator calculation requirements;

[0036] A list of wall-penetrating segments is calculated based on the boundary restriction information, the indoor positioning trajectory data, and the obstacle weight map.

[0037] The trajectory evaluation index is calculated based on the list of wall-penetrating segments and the index calculation requirements.

[0038] In one possible implementation, the step of analyzing the trajectory evaluation indicators based on a preset compliance verification process to generate a compliance result for the indoor positioning trajectory data includes:

[0039] Determine whether the maximum single-anoma average pixel score and the maximum single-anoma score in the trajectory evaluation index are both less than their respective first and second set thresholds.

[0040] If any one of them is not met, then it is determined whether the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds.

[0041] If the overall trajectory score and the average trajectory score in the trajectory evaluation index are not greater than their respective third and fourth set thresholds, the trajectory is deemed compliant.

[0042] If the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds, then the trajectory is determined to be non-compliant.

[0043] If both conditions are met, it is determined that the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are not both greater than their respective third and fifth set thresholds.

[0044] If the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are both greater than their respective third and fifth set thresholds, the trajectory is determined to be non-compliant.

[0045] If the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are not both greater than their respective third and fifth set thresholds, then the trajectory is deemed compliant.

[0046] In a second aspect of this application, a method and apparatus for determining compliance of wall crossing and boundary intrusion based on scene graph and trajectory analysis is provided. The apparatus includes:

[0047] The area data acquisition module is used to acquire building structure information and indoor positioning trajectory data of the area to be judged;

[0048] An obstacle image generation module is used to generate an overall obstacle map based on the building structure information;

[0049] The boundary index generation module is used to perform edge detection and feature fusion processing on the overall obstacle map to generate boundary point coordinate indexes;

[0050] The obstacle weight generation module is used to perform iterative convolution operations on the boundary point coordinate index and the overall obstacle map to generate an obstacle weight map.

[0051] The evaluation index calculation module is used to calculate the trajectory evaluation index based on the indoor positioning trajectory data and the obstacle weight map;

[0052] The verification result generation module is used to analyze the trajectory evaluation indicators based on a preset compliance verification process and generate compliance results for the indoor positioning trajectory data.

[0053] In a third aspect of this application, an electronic device is provided. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described above.

[0054] In a fourth aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the method according to the first aspect of this application.

[0055] The wall-crossing and boundary-crossing compliance judgment method based on scene graph and trajectory analysis provided in this application innovatively utilizes scene graph and iterative convolution technology to construct a weight graph that can accurately reflect spatial access constraints by integrating building structure information and positioning trajectory data. This effectively overcomes the problem of trajectory misjudgment caused by positioning errors in the prior art. It does not rely on complex mathematical models and large-scale data training, significantly reducing the computational complexity and resource consumption of the algorithm, and improving the recognition efficiency and judgment reliability of abnormal wall-crossing and boundary-crossing segments. By introducing trajectory evaluation indicators and preset compliance verification processes, it can comprehensively evaluate trajectory compliance, significantly enhance the detection capability of abnormal behavior in indoor positioning trajectories, and achieve the goal of improving the accuracy of indoor positioning trajectory analysis.

[0056] It should be understood that the description in the Summary Section is not intended to limit the key or essential features of the embodiments of this application, nor is it intended to restrict the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0057] The above and other features, advantages, and aspects of the embodiments of this application will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0058] Figure 1 This is a flowchart of a method for determining compliance of wall crossing and boundary crossing based on scene graph and trajectory analysis according to an embodiment of this application;

[0059] Figure 2 For the overall obstacle diagram according to this application;

[0060] Figure 3 A diagram of the obstacle structure according to this application;

[0061] Figure 4 A partial obstacle structure diagram according to this application;

[0062] Figure 5 This is a block diagram of a wall-crossing and boundary-crossing compliance judgment device based on scene graph and trajectory analysis according to an embodiment of this application;

[0063] Figure 6 This is a schematic diagram of the structure of a terminal device or server suitable for implementing the embodiments of this application. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0065] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0066] Figure 1 A flowchart is shown for a method for determining compliance of wall crossing and boundary crossing based on scene graph and trajectory analysis according to an embodiment of the present disclosure.

[0067] like Figure 1 As shown, the main process of this method is described below (steps S101 to S106):

[0068] Step S101: Obtain building structure information and indoor positioning trajectory data for the area to be judged.

[0069] In some embodiments, the area to be judged is a building area that needs to be judged through walls, such as a shopping mall, residential building or office building, etc. The building structure information is the building scene drawing of the area to be judged, as well as the planning, annotation and other data corresponding to the building scene drawing. The indoor positioning trajectory data is the trajectory information generated during the movement when positioning is performed indoors.

[0070] Step S102: Generate an overall obstacle map based on the building structure information.

[0071] For step S102, the building walls and building boundaries are determined based on the building structure information; a rigid wall map is generated based on the building walls and building boundaries; a binarization allocation scheme is obtained, wherein the binarization allocation scheme sets the area outside the building walls and building boundaries as the first pixel value and the passable area as the second pixel value; the rigid wall map is binarized based on the binarization allocation scheme to generate an overall obstacle map.

[0072] In some embodiments, the images corresponding to the building walls and building boundaries in the input building structure information are extracted as rigid wall images. The images corresponding to the building walls and building boundaries are from the building scene drawings, and only the walls and boundaries are retained. If the images corresponding to the building walls and building boundaries are separate, the images corresponding to the building walls and building boundaries need to be synthesized to obtain the final rigid wall image.

[0073] rigid walls Figure 2 Value transformation as follows Figure 2 The overall obstacle map shown is binarized according to the set binarization allocation scheme, where the area outside the building walls and boundaries is set as the first pixel value, and the passable area is set as the second pixel value. Figure 2 In the diagram, the white areas represent obstacles such as walls and load-bearing columns, i.e., building wall areas, with pixel values ​​set to 1. The remaining movable areas have values ​​of 0. Since the overall obstacle map is binarized, pixels with a value of 1 are white, and pixels with a value of 0 are black. In addition to distinguishing between them, setting the pixel value of building walls to 1 will increase the value of subsequent convolution operations, making the more walls the trajectory passes through, the greater the weight.

[0074] Step S103: Perform edge detection and feature fusion processing on the overall obstacle map to generate boundary point coordinate indexes.

[0075] For step S103, the target image processing direction is obtained; the overall obstacle map is deconvolved based on the convolution kernel of the target image processing direction to obtain a multi-directional feature map; the multi-directional feature map is spatially aligned and fused to generate a processing feature map; the processing feature map is pooled and a Boolean matrix of marked boundary points is generated by filtering through a preset threshold; the Boolean matrix is ​​used to extract data to obtain the boundary point coordinate index.

[0076] In some embodiments, edge detection is performed on the overall obstacle map. During processing, the target image processing direction is set to nine. Nine 3×3 convolutional kernels, sensitive to edge features in different directions, are used to deconvolve the overall obstacle map, reflecting the response preference to pixel changes in specific directions. Deconvolution maps each pixel of the input image to a 3×3 feature window, effectively expanding the local features of a single pixel into spatial distribution information. Differential padding and cropping ensure spatial alignment of the feature maps in the nine directions. Then, multi-directional features are fused to obtain a robust processed feature map that enhances boundary information. Max pooling is performed on the fused feature map, and regions with strong responses are selected by thresholding. Finally, a Boolean matrix marking boundary points is obtained, and the boundary point coordinate indices are extracted from the Boolean matrix. It should be noted that the default threshold is set to 0.5, but it can be modified to a suitable value according to actual conditions. The specific range of threshold modification is not limited here.

[0077] Step S104: Perform iterative convolution operations on the boundary point coordinate indices and the overall obstacle map to generate an obstacle weight map.

[0078] For step S104, the boundary position pixels in the overall obstacle map are weighted based on the boundary point coordinate index to obtain an initial weighted map; the initial weighted map is convolved using a preset convolution kernel to generate a convolution result; the convolution result is incremented by 1 and its reciprocal is taken to obtain a first intermediate feature map; the overall obstacle map is convolved to generate a neighborhood summation mapping, and the pixels at the boundary point coordinate index positions are normalized to obtain a second intermediate feature map; an initial weight map is generated based on the first intermediate feature map, the second intermediate feature map, the overall obstacle map, and a preset number of iterations; the initial weight map is post-processed at the boundary to generate an obstacle weight map.

[0079] In some embodiments, a full-1 convolution kernel is used to perform multiple iterative convolutions on the overall obstacle map to locate the boundary point coordinates. The boundary position pixels in the overall obstacle map are then weighted using a coefficient α to obtain an initial weighted map. The coefficient α is set to 0.7 and can be modified according to the actual situation. Due to the large number of iterations, a coefficient α greater than 1 will cause the boundary weights to increase too quickly, affecting subsequent score judgments. A 3×3 full-1 convolution kernel k1 is used to convolve the initial weighted map to obtain the convolution result. The first intermediate feature map is obtained by adding 1 to the convolution result and taking its reciprocal. This processing scheme can amplify the relative weight of fine boundary regions. Since the value of fine boundaries is small after convolution, taking the reciprocal will result in a larger value, thus solving the problem of excessively low weight of fine boundaries in feature fusion. Here, fine boundaries refer to boundaries in the image that are extremely narrow, usually only 1-2 pixels wide, and have weak edge signals. When convolving, the sum of local pixels in the fine boundary region will be significantly smaller than that of the coarse boundary because the fine boundary occupies only a few pixels, resulting in a lower sum. The smaller, finer boundary values ​​are amplified in relative weight after reciprocal calculation, thus preventing them from being suppressed by the strong signals of coarser boundaries in subsequent fusion. Then, a full-1 convolution kernel with a size varying with the iteration count is generated. This kernel is convolved with the overall obstacle map and multiplied by the overall obstacle map to obtain a neighborhood summation map, reflecting the overall intensity of the region surrounding each pixel. Simultaneously, the first intermediate feature map is convolved with the full-1 convolution kernel, and the result is multiplied by the overall obstacle map. The mapping is then used to normalize the coordinate indices of specific boundary points, resulting in the second intermediate feature map. This normalization process avoids the problem of excessively large values ​​in edge regions due to fewer neighboring pixels, making the boundary features more stable. Finally, the first and second intermediate feature maps are weighted and fused using a coefficient β: first intermediate feature map * second intermediate feature map * β, where β is set to 2 (or other values ​​depending on actual needs). This fusion is then multiplied by the overall obstacle map to obtain the initial weight map for the next iteration. Finally, boundary post-processing is applied to the initial weight map to obtain the final obstacle weight map.

[0080] Furthermore, the initial weight map is subjected to boundary post-processing to generate an obstacle weight map, including: determining the nearest neighbor region within a preset pixel length range inside and outside the boundary based on the rigid wall map; performing interference filtering on the corresponding positions in the nearest neighbor region and the initial weight map; setting the weight value in the preset far boundary region of the initial weight map as the maximum weight value in the initial weight map, thus obtaining the final obstacle weight map.

[0081] After iteration, the inner boundary pixel length is set to 3 and the outer boundary pixel length to 15. A convolution kernel of all 1s is used to convolve the rigid wall image. The intersection of the result with the range yields the inner and outer nearest neighbor regions of the boundary. These inner and outer nearest neighbor regions are merged, and the position with a value of 1 is found. The corresponding weight in the initial weight map is reset to 0, filtering out interference weights near the boundary. The weights of the regions outside the outer boundary but not within the outer boundary pixel length range in the initial weight map are set to the maximum value of the initial weight map, highlighting the weights of the far boundary regions. After processing, the obstacle weight map is obtained.

[0082] Step S105: Calculate trajectory evaluation index based on indoor positioning trajectory data and obstacle weight map.

[0083] For step S105, obtain boundary restriction information and indicator calculation requirements; calculate a wall penetration segment list based on boundary restriction information, indoor positioning trajectory data and obstacle weight map; calculate trajectory evaluation indicators based on the wall penetration segment list and indicator calculation requirements.

[0084] In some embodiments, the indicator calculation requirements include the average pixel score of the trajectory, the maximum single anomaly score, the average pixel score of the maximum single anomaly, the overall trajectory score, and the average trajectory score. Relevant indicators are calculated using an obstacle weight map and the trajectory. First, the trajectory coordinates and the obstacle weight map are compared to obtain a list of all wall-penetrating segments. When calculating the list of wall-penetrating segments, considering practical considerations, boundary restrictions are set. For example, there are no restrictions on the first floor's boundary crossing; for other floors, the pixel score outside the boundary is set to the maximum weight in the current obstacle weight map. For the second floor and above, considering that users may walk along the boundary scaffolding, a certain range of zero-value weights is allowed. The specific range restriction can be set by specifying parameters when building the obstacle weight map. The weights of all wall-penetrating segments are added together to obtain the total weight and the total length of the wall-penetrating segments. The total weight is divided by the total length to obtain the average weight. The maximum weight among all wall-penetrating segments is used as the maximum single anomaly score, and divided by the segment length to obtain the maximum single anomaly average score. The total weight is divided by the total trajectory length to obtain the average pixel score of this trajectory. The total weight is used as the overall trajectory score, and the average weight is used as the average trajectory score.

[0085] Step S106: Analyze the trajectory evaluation indicators based on the preset compliance verification process to generate compliance results for indoor positioning trajectory data.

[0086] For step S106, it is determined whether the maximum single-abnormality average pixel score and the maximum single-abnormality score in the trajectory evaluation index are both less than their respective first and second set thresholds. If either one is not satisfied, it is determined whether the overall trajectory score and the average trajectory score in the trajectory evaluation index are both less than their respective third and fourth set thresholds. If both the overall trajectory score and the average trajectory score in the trajectory evaluation index are less than their respective third and fourth set thresholds, the trajectory is deemed compliant. If both the overall trajectory score and the average trajectory score in the trajectory evaluation index are less than their respective third and fourth set thresholds, the trajectory is deemed non-compliant. If both are satisfied, it is determined whether the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are both greater than their respective third and fifth set thresholds. If both the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are greater than their respective third and fifth set thresholds, the trajectory is deemed non-compliant. If neither the overall trajectory score nor the average trajectory pixel score in the trajectory evaluation index is greater than their respective third and fifth set thresholds, the trajectory is deemed compliant.

[0087] In some embodiments, it is first determined whether the maximum single-abnormal average pixel score and the maximum single-abnormal score are respectively less than a set threshold, that is, whether the maximum single-abnormal average pixel score is less than a first set threshold and whether the maximum single-abnormal score is respectively less than a second set threshold. If either one is not satisfied, the second judgment logic is executed; otherwise, the third judgment logic is executed.

[0088] The second judgment logic is to determine whether the overall trajectory score and the average trajectory score are greater than a certain threshold. Specifically, it is to determine whether the overall trajectory score in the trajectory evaluation index is not greater than the third set threshold and whether the average trajectory score is not greater than the fourth set threshold. If the overall trajectory score is not greater than the third set threshold and the average trajectory score is not greater than the fourth set threshold, then it is considered that a proper wall crossing has occurred in a long trajectory, and the trajectory is judged to be compliant. If the overall trajectory score is greater than the third set threshold and the average trajectory score is not greater than the fourth set threshold, or the overall trajectory score is not greater than the third set threshold and the average trajectory score is greater than the fourth set threshold, or the overall trajectory score is greater than the third set threshold and the average trajectory score is also greater than the fourth set threshold, then the trajectory is judged to be non-compliant.

[0089] The third judgment logic is to determine whether the overall trajectory score and the average trajectory pixel score are greater than the set thresholds, that is, whether the overall trajectory score in the trajectory evaluation index is greater than the third set threshold and whether the average trajectory pixel score is greater than the fifth set threshold. If the overall trajectory score in the trajectory evaluation index is greater than the third set threshold and the average trajectory pixel score is greater than the fifth set threshold, then the trajectory is determined to be non-compliant. If the overall trajectory score in the trajectory evaluation index is not greater than the third set threshold and the average trajectory pixel score is greater than the fifth set threshold, or the overall trajectory score in the trajectory evaluation index is greater than the third set threshold and the average trajectory pixel score is not greater than the fifth set threshold, or the overall trajectory score in the trajectory evaluation index is not greater than the third set threshold and the average trajectory pixel score is not greater than the fifth set threshold, then the trajectory is determined to be compliant.

[0090] It should be noted that the thresholds can be selected by the user, and the thresholds for the five indicators are different. Here, the thresholds are set as follows: the maximum single abnormality average pixel score (the first set threshold) is 0.2, the maximum single abnormality score threshold (the second set threshold) is 0.55, the overall trajectory score threshold (the third set threshold) is 1.5, the average trajectory score threshold (the fourth set threshold) is 0.15, and the trajectory pixel average score threshold (the fifth set threshold) is 0.03.

[0091] According to the embodiments of this disclosure, the following technical effects are achieved: By integrating building structure information and positioning trajectory data, a weighted graph that can accurately reflect spatial access constraints is innovatively constructed using scene graphs and iterative convolution techniques. This effectively overcomes the problem of trajectory misjudgment caused by positioning errors in the prior art. It does not rely on complex mathematical models and large-scale data training, significantly reducing the computational complexity and resource consumption of the algorithm, and improving the recognition efficiency and judgment reliability of abnormal wall-crossing and boundary-crossing segments. By introducing trajectory evaluation indicators and preset compliance verification processes, trajectory compliance can be comprehensively evaluated, significantly enhancing the ability to detect abnormal behavior in indoor positioning trajectories, and achieving the goal of improving the accuracy of indoor positioning trajectory analysis.

[0092] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0093] The above is an introduction to the method embodiments. The following describes the solution described in this application through device embodiments.

[0094] Figure 5A block diagram of a wall-crossing and boundary-crossing compliance judgment device based on scene graph and trajectory analysis according to an embodiment of this application is shown, as follows: Figure 5 The following are included:

[0095] The area data acquisition module 201 is used to acquire building structure information and indoor positioning trajectory data of the area to be judged.

[0096] The obstacle image generation module 202 is used to generate an overall obstacle map based on building structure information;

[0097] The boundary index generation module 203 is used to perform edge detection and feature fusion processing on the overall obstacle map and generate boundary point coordinate indexes.

[0098] The obstacle weight generation module 204 is used to perform iterative convolution operations on the boundary point coordinate index and the overall obstacle map to generate an obstacle weight map.

[0099] The evaluation index calculation module 205 is used to calculate trajectory evaluation indexes based on indoor positioning trajectory data and obstacle weight map.

[0100] The verification result generation module 206 is used to analyze the trajectory evaluation indicators based on the preset compliance verification process and generate compliance results for indoor positioning trajectory data.

[0101] As an optional implementation of this embodiment, the obstacle image generation module 202 is specifically used to determine the building walls and building boundaries based on the building structure information; generate a rigid wall image based on the building walls and building boundaries; obtain a binarization allocation scheme, wherein the binarization allocation scheme sets the area outside the building walls and building boundaries as the first pixel value and the passable area as the second pixel value; and perform binarization processing on the rigid wall image based on the binarization allocation scheme to generate an overall obstacle image.

[0102] As an optional implementation of this embodiment, the boundary index generation module 203 is specifically used to obtain the target image processing direction; perform deconvolution operation on the overall obstacle map based on the convolution kernel of the target image processing direction to obtain a multi-directional feature map; perform spatial alignment and summation fusion on the multi-directional feature map to generate a processing feature map; perform pooling processing on the processing feature map and generate a Boolean matrix for marking boundary points by filtering through a preset threshold; and extract data from the Boolean matrix to obtain the boundary point coordinate index.

[0103] As an optional implementation of this embodiment, the obstacle weight generation module 204 includes:

[0104] The initial image generation module is used to weight the boundary position pixels in the overall obstacle map based on the boundary point coordinate index to obtain the initial weighted image.

[0105] The processing result generation module is used to perform convolution processing on the initial weighted image using a preset convolution kernel to generate the convolution processing result.

[0106] The first feature generation module is used to add 1 to the convolution result and take the reciprocal to obtain the first intermediate feature map;

[0107] The second feature generation module is used to perform convolution on the overall obstacle map to generate a neighborhood summation mapping, and to normalize the pixels at the coordinate index positions of the boundary points to obtain the second intermediate feature map.

[0108] The weighted image generation module is used to generate an initial weighted image based on the first intermediate feature map, the second intermediate feature map, the overall obstacle map, and a preset number of iterations.

[0109] The obstacle image generation module is used to perform boundary post-processing on the initial weight map to generate an obstacle weight map.

[0110] In this optional embodiment, the obstacle image generation module is specifically used to determine the nearest neighbor region within a preset pixel length range inside and outside the boundary based on the rigid wall map; to perform interference filtering processing on the corresponding positions of the nearest neighbor region and the initial weight map; and to set the weight value of the region in the preset far boundary region in the initial weight map as the maximum weight value in the initial weight map to obtain the final obstacle weight map.

[0111] As an optional implementation of this embodiment, the evaluation index calculation module 205 is specifically used to obtain boundary restriction information and index calculation requirements; calculate a wall penetration segment list based on the boundary restriction information, indoor positioning trajectory data and obstacle weight map; and calculate trajectory evaluation index based on the wall penetration segment list and index calculation requirements.

[0112] As an optional implementation of this embodiment, the verification result generation module 206 is specifically used to determine whether the maximum single-abnormality average pixel score and the maximum single-abnormality score in the trajectory evaluation index are both less than their respective first and second set thresholds; if one of them is not satisfied, then it is determined whether the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds; if the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds, then the trajectory is determined to be compliant; if the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds, then the trajectory is determined to be non-compliant; if both of them are satisfied, then it is determined that the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are not both greater than their respective third and fifth set thresholds; if the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are both greater than their respective third and fifth set thresholds, then the trajectory is determined to be non-compliant; if the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are not both greater than their respective third and fifth set thresholds, then the trajectory is determined to be compliant.

[0113] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0114] Figure 6 A schematic diagram of the structure of a terminal device or server suitable for implementing the embodiments of this application is shown.

[0115] like Figure 6 As shown, the terminal device or server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 302 or a program loaded from storage section 508 into random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the terminal device or server. The CPU 301, ROM 302, and RAM 303 are interconnected via bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0116] The following components are connected to I / O interface 305: an input section 306 including a keyboard, mouse, etc.; an output section 307 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN card, modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to I / O interface 305 as needed. A removable medium 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 310 as needed so that computer programs read from it can be installed into storage section 308 as needed.

[0117] Specifically, according to embodiments of this application, the above method flow steps can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a machine-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by central processing unit (CPU) 301, it performs the functions defined in the system of this application.

[0118] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0119] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0120] The units or modules described in the embodiments of this application can be implemented in software or hardware. The described units or modules can also be located in a processor. The names of these units or modules do not, in certain circumstances, constitute a limitation on the unit or module itself.

[0121] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium stores one or more programs that, when used by one or more processors, execute the methods described in this application.

[0122] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. A method for judging compliance of wall crossing and boundary violation based on scene graph and trajectory analysis, characterized in that, include: Obtain building structure information and indoor positioning trajectory data for the area to be judged; A comprehensive obstacle map is generated based on the aforementioned building structure information; Edge detection and feature fusion are performed on the overall obstacle map to generate boundary point coordinate indices; Iterative convolution operations are performed on the boundary point coordinate indices and the overall obstacle map to generate an obstacle weight map, including: Based on the boundary point coordinate index, the boundary position pixels in the overall obstacle map are weighted to obtain an initial weighted map; The initial weighted image is convolved using a preset convolution kernel to generate a convolution result. Add 1 to the convolution result and take the reciprocal to obtain the first intermediate feature map; The overall obstacle map is convolved to generate a neighborhood summation mapping, and the pixels at the coordinate index positions of the boundary points are normalized to obtain a second intermediate feature map. An initial weight map is generated based on the first intermediate feature map, the second intermediate feature map, the overall obstacle map, and a preset number of iterations. The initial weight map is subjected to boundary post-processing to generate an obstacle weight map; The trajectory evaluation index is calculated based on the indoor positioning trajectory data and the obstacle weight map; The trajectory evaluation indicators are analyzed based on a preset compliance verification process to generate compliance results for the indoor positioning trajectory data.

2. The method according to claim 1, characterized in that, The generation of the overall obstacle map based on the building structure information includes: The building walls and building boundaries are determined based on the aforementioned building structure information; A rigid wall diagram is generated based on the building walls and the building boundaries; Obtain a binarization allocation scheme, wherein the binarization allocation scheme sets the area outside the building walls and building boundaries as the first pixel value and the passable area as the second pixel value; The rigid wall map is binarized based on the aforementioned binarization allocation scheme to generate an overall obstacle map.

3. The method according to claim 1, characterized in that, The step of performing edge detection and feature fusion processing on the overall obstacle map to generate boundary point coordinate indices includes: Obtain the target image processing direction; The overall obstacle map is deconvolved based on the convolution kernels in the target image processing direction to obtain a multi-directional feature map; The multi-directional feature maps are spatially aligned and fused to generate a processed feature map; The processed feature map is pooled, and a Boolean matrix of marked boundary points is generated by filtering through a preset threshold. Data extraction is performed on the Boolean matrix to obtain the coordinate indices of the boundary points.

4. The method according to claim 3, characterized in that, The step of performing boundary post-processing on the initial weight map to generate an obstacle weight map includes: Based on the rigid wall diagram, determine the nearest neighbor region within a preset pixel length range inside and outside the boundary; Interference filtering is performed on the neighboring regions and the corresponding positions in the initial weight map; The weight values ​​in the preset far boundary region of the initial weight map are set to the maximum weight values ​​in the initial weight map to obtain the final obstacle weight map.

5. The method according to claim 1, characterized in that, The trajectory evaluation index calculated based on the indoor positioning trajectory data and the obstacle weight map includes: Obtain outbound limit information and indicator calculation requirements; A list of wall-penetrating segments is calculated based on the boundary restriction information, the indoor positioning trajectory data, and the obstacle weight map. The trajectory evaluation index is calculated based on the list of wall-penetrating segments and the index calculation requirements.

6. The method according to claim 1, characterized in that, The process of analyzing the trajectory evaluation indicators based on a preset compliance verification procedure to generate compliance results for the indoor positioning trajectory data includes: Determine whether the maximum single-anoma average pixel score and the maximum single-anoma score in the trajectory evaluation index are both less than their respective first and second set thresholds. If any one of them is not met, then it is determined whether the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds. If the overall trajectory score and the average trajectory score in the trajectory evaluation index are not greater than their respective third and fourth set thresholds, the trajectory is deemed compliant. If the overall trajectory score and the average trajectory score in the trajectory evaluation index are both not greater than their respective third and fourth set thresholds, then the trajectory is determined to be non-compliant. If both conditions are met, it is determined that the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are not both greater than their respective third and fifth set thresholds. If the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are both greater than their respective third and fifth set thresholds, the trajectory is determined to be non-compliant. If the overall trajectory score and the average trajectory pixel score in the trajectory evaluation index are not both greater than their respective third and fifth set thresholds, then the trajectory is deemed compliant.

7. A method and device for judging compliance of wall crossing and boundary violation based on scene graph and trajectory analysis, characterized in that, include: The area data acquisition module is used to acquire building structure information and indoor positioning trajectory data of the area to be judged; An obstacle image generation module is used to generate an overall obstacle map based on the building structure information; The boundary index generation module is used to perform edge detection and feature fusion processing on the overall obstacle map to generate boundary point coordinate indexes; The obstacle weight generation module is used to perform iterative convolution operations on the boundary point coordinate indices and the overall obstacle map to generate an obstacle weight map, including: Based on the boundary point coordinate index, the boundary position pixels in the overall obstacle map are weighted to obtain an initial weighted map; The initial weighted image is convolved using a preset convolution kernel to generate a convolution result. Add 1 to the convolution result and take the reciprocal to obtain the first intermediate feature map; The overall obstacle map is convolved to generate a neighborhood summation mapping, and the pixels at the coordinate index positions of the boundary points are normalized to obtain a second intermediate feature map. An initial weight map is generated based on the first intermediate feature map, the second intermediate feature map, the overall obstacle map, and a preset number of iterations. The initial weight map is subjected to boundary post-processing to generate an obstacle weight map; The evaluation index calculation module is used to calculate the trajectory evaluation index based on the indoor positioning trajectory data and the obstacle weight map; The verification result generation module is used to analyze the trajectory evaluation indicators based on a preset compliance verification process and generate compliance results for the indoor positioning trajectory data.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.

Images