Building intelligent positioning system
By collecting facial and motion data in the building intelligent positioning system, a tracking and positioning model is established. By utilizing the mobile execution module and positioning track, the problem that existing technologies can only locate at the entrance is solved, realizing multi-location tracking and positioning, and improving security and image continuity.
Patent Information
- Application Number
- CN202410989563.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-23
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing building intelligent positioning systems can only locate individuals at the entrance, and cannot identify different people at multiple locations, resulting in insufficient security and image continuity.
By collecting facial and motion data, a tracking and positioning model is established. Positioning is performed in multiple locations using positioning cameras and motion execution modules. By combining image processing and computer vision technologies, facial and motion data can be labeled, preprocessed, and model trained. Positioning tracks and rails are used to move the camera, achieving multi-location tracking and positioning.
It enables effective tracking and positioning of different people in multiple locations, improving security and image continuity, and enhancing positioning performance.
Smart Images

Figure CN118982579A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of building intelligent positioning technology, and more specifically, to a building intelligent positioning system. Background Art
[0002] With the development of modern technology, multiple cameras are usually installed in offices and other places to locate different visitors, thereby providing security and facilitating management.
[0003] The document with the prior art publication number CN112911515B provides a building intelligent positioning system. The device is provided with a central controller, a monitoring camera, an indicator light, a wireless transmitter and a mobile phone. The monitoring camera, the indicator light and the wireless transmitter are all electrically connected to the central controller. The wireless transmitter is wirelessly connected to the mobile phone via wireless WiFi. After the participants in the meeting enter the building, the mobile phone on their body will connect to and locate the wireless transmitter in the building. The wireless transmitter will send a signal to the central controller, indicating that a participant has entered the building. Then the monitoring camera will perform facial recognition on the participant and send the information to the central controller. The central controller will compare the obtained personnel information with the pre-stored information to determine where the person should go. Then the central controller controls the indicator lights at each intersection to help the person go to the destination.
[0004] Although the above-mentioned prior art solution can achieve the relevant beneficial effects through the structure of the prior art, it still has the following defects: the device obtains and locates different people through a camera fixed on one side, and usually can only locate at the entrance, and cannot identify different situations.
[0005] With respect to the above-mentioned related technologies, the inventor believes that when positioning different persons, it is necessary to position different persons at multiple positions, so as to effectively grasp their movement trends and improve safety and image continuity.
[0006] In view of this, we propose a building intelligent positioning system. Summary of the invention
[0007] 1. Technical issues to be solved
[0008] The purpose of the present application is to provide a method for a building intelligent positioning system, which solves the technical problem that the device in the above-mentioned background technology obtains and locates different people through a camera fixed on one side, usually can only locate at the entrance, and cannot identify different situations, thereby achieving a technical effect.
[0009] 2. Technical solution
[0010] The technical solution of this application provides a building intelligent positioning system, including
[0011] The face and motion data collection module collects facial expression and motion data through positioning cameras;
[0012] The face and motion data marking module marks suspicious facial expressions and motion data;
[0013] The face and motion data preprocessing module processes the face and motion data to improve the quality of the face and motion data in order to establish a model;
[0014] Establish a tracking and positioning model, and use face and motion data to build a suitable model;
[0015] The tracking and positioning model training and tuning module uses labeled data sets to train the model and iteratively trains the model, adjusting hyperparameters to improve performance;
[0016] Save the tracking and positioning model and apply the module to save and apply the trained tracking and positioning model;
[0017] The mobile execution module can drive the positioning camera 4 to perform further positioning.
[0018] By adopting the above technical solution and establishing a tracking model, people can be analyzed after their entrance is obtained and followed up. Different people can be located in multiple positions, so that their movement trends can be effectively grasped, improving safety and image continuity.
[0019] As an optional solution of the technical solution of the present application document, the face and action data marking module uses manual inspection of each image or video frame, and then manually draws a bounding box or mark through desktop graphics editing software;
[0020] Use computer vision tools and algorithms to automatically detect faces and provide preliminary bounding boxes. These preliminary markings are then manually reviewed and corrected to ensure accuracy.
[0021] Use the key points of the face to mark, such as the positions of the eyes, nose, mouth, etc.
[0022] As an optional solution of the technical solution of the present application document, the face and motion data preprocessing module uses image processing technology to improve image quality, such as denoising, sharpening or brightness adjustment, adjusts the image size to make it have the same size so that the model can process fixed-size inputs, increases the diversity of data by applying random transformations, improves the generalization ability of the model, normalizes the image data, and scales the pixel values to a specific range, usually [0,1] or [-1,1]. During training, additional data enhancement techniques can be applied, such as random cropping, color enhancement, etc., to increase the diversity of the data.
[0023] As an optional solution of the technical solution of the present application document, the tracking and positioning model training and tuning module divides the data into a training set, a validation set and a test set, with most of the data used for training, a part for validating the performance of the model, and a part for the final test; when performing tuning, an appropriate loss function is selected, which will measure the performance of the model and the accuracy of the prediction, and a stochastic gradient descent optimization algorithm is selected to perform forward propagation and back propagation on the training data to update the weights and parameters of the model, and the model performance is regularly evaluated on the validation set to detect overfitting or underfitting; hyperparameter adjustment is performed, and the hyperparameters include learning rate, batch size, number of layers, number of neurons, regularization strength, etc., and grid search, random search or experience-based methods are used to adjust the hyperparameters.
[0024] As an optional solution of the technical solution of the present application document, the mobile execution module includes a positioning track, and the positioning track can be set at entrances and corridors in different directions;
[0025] Positioning camera, which can obtain and locate human images;
[0026] Track connection: positioning tracks in different directions can be connected by track connection;
[0027] The mobile component can drive the positioning camera to follow the moving position, so that the image can be positioned multiple times continuously.
[0028] 3. Beneficial effects
[0029] One or more technical solutions provided in the technical solution of this application have at least the following technical effects or advantages:
[0030] 1. This application establishes a tracking model, which can analyze people after they enter the scene and then follow up. It can locate different people in multiple locations, so as to effectively grasp their movement trends and improve safety and image continuity.
[0031] 2. This application can connect to positioning tracks in different directions through the setting of the connection, so that the camera can be driven to move in different directions, so that the same person can be tracked and located, thereby improving the positioning effect.
[0032] 3. This application can actively drive the camera to move and turn vertically through the setting of mobile components. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 This is a schematic diagram of the overall structure of a building intelligent positioning system disclosed in a preferred embodiment of the present application;
[0034] Figure 2 A bottom view of the overall structure of a building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0035] Figure 3 This is an expanded view of the overall structure of a building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0036] Figure 4 This is an enlarged view of the structure of location A of a building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0037] Figure 5 A schematic diagram of the relationship between the positioning track and the track connection structure of the building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0038] Figure 6 A schematic diagram of the structure of a mobile component of a building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0039] Figure 7 An exploded diagram of the structure of a mobile component of a building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0040] Figure 8 A schematic diagram of the internal structure of a mobile block of a building intelligent positioning device disclosed in a preferred embodiment of the present application;
[0041] Explanation of the numbers in the figure: 1. Positioning track; 11. Track connection; 101. Rotating groove; 102. Rotating block; 103. Machine cavity; 104. First motor; 2. Fixed block; 3. Building wall; 4. Positioning camera; 5. Moving assembly; 51. Moving block; 52. Roller; 53. Connecting block; 54. Limit block; 55. Limit groove; 56. Gear groove; 57. Transmission gear; 58. Internal gear; 501. Mechanical groove; 503. Rotating shaft; 504. External gear; 505. Driving gear; 506. Second motor. DETAILED DESCRIPTION
[0042] The present application is further described in detail below in conjunction with the accompanying drawings.
[0043] Reference Figure 1 The present application embodiment provides a building intelligent positioning system, including
[0044] The face and motion data collection module collects the face expression and motion data through the positioning camera 4;
[0045] The face and motion data marking module marks suspicious facial expressions and motion data;
[0046] The face and motion data preprocessing module processes the face and motion data to improve the quality of the face and motion data in order to establish a model;
[0047] Establish a tracking and positioning model, and use face and motion data to build a suitable model;
[0048] The tracking and positioning model training and tuning module uses labeled data sets to train the model and iteratively trains the model, adjusting hyperparameters to improve performance;
[0049] Save the tracking and positioning model and apply the module to save and apply the trained tracking and positioning model;
[0050] The mobile execution module can drive the positioning camera for further positioning.
[0051] By building a tracking model, we can analyze people after obtaining their entrances and then follow up. We can locate different people in multiple locations, so that we can effectively grasp their movement trends and improve safety and image continuity.
[0052] Reference Figure 1 and Figure 2 ,The embodiment of the present application provides a building intelligent positioning system ,with a face and motion data marking module that utilizes manual inspection of each image or video frame and then manually draws a bounding box or marking through desktop ,graphics editing software;
[0053] Use computer vision tools and algorithms to automatically detect faces and provide preliminary bounding boxes. These preliminary markings are then manually reviewed and corrected to ensure accuracy.
[0054] Use the key points of the face to mark, such as the positions of the eyes, nose, mouth, etc.
[0055] The face and action data preprocessing module uses image processing techniques to improve image quality, such as denoising, sharpening or brightness adjustment, adjusts the image size to make them of the same size so that the model can handle fixed-size inputs, increases the diversity of data by applying random transformations such as rotation, flipping, translation, and scaling, improves the generalization ability of the model, normalizes the image data, and scales the pixel values to a specific range, usually [0,1] or [-1,1]. During training, additional data enhancement techniques such as random cropping, color enhancement, etc. can be applied to increase the diversity of the data.
[0056] The tracking and positioning model training and tuning module divides the data into training set, validation set and test set. Most of the data is used for training, part of it is used to verify the performance of the model, and part of it is used for final testing. When tuning, select an appropriate loss function, which will measure the performance of the model and the accuracy of the prediction. Select the stochastic gradient descent SGD optimization algorithm, perform forward propagation and back propagation on the training data to update the weights and parameters of the model, and regularly evaluate the model performance on the validation set to detect overfitting or underfitting. Perform hyperparameter adjustment, which includes learning rate, batch size, number of layers, number of neurons, regularization strength, etc. Use grid search, random search or experience-based methods to adjust hyperparameters.
[0057] Mobile execution modules include
[0058] Positioning track 1, positioning track 1 can be set at entrances and corridors in different directions;
[0059] Positioning camera 4, which can obtain and locate the image of a person;
[0060] A connecting rail 11, which can be used to connect the positioning rails 1 in different directions;
[0061] The moving component 5 can drive the positioning camera 4 to follow the moving position, so that the image can be positioned multiple times continuously.
[0062] Reference Figure 2 and Figure 3 The present application embodiment provides a building intelligent positioning system, including
[0063] A positioning track 1, wherein the positioning track 1 is fixedly arranged on an outer wall of a building wall 3 through a fixing block 2;
[0064] The positioning track 1 can be arranged along the building wall 3 and arranged in a T-shaped structure, or other vertical tracks;
[0065] Different positioning rails 1 are connected and switched via connecting rails 11 .
[0066] The positioning camera 4 is slidably arranged on the outer wall of the positioning track 1 through a moving component 5 .
[0067] Reference Figure 3 , Figure 4 and Figure 5 The embodiment of the present application provides a building intelligent positioning system, wherein a rotation groove 101 is provided on one side of the positioning track 1, a rotation block 102 is rotatably arranged inside the rotation groove 101, and one end of the rotation block 102 extends out of the rotation groove 101 to the outside and is connected and fixed to the connecting rail 11;
[0068] An engine chamber 103 is provided on one side of the rotating tank 101 . A first motor 104 is disposed inside the engine chamber 103 . The output shaft of the first motor 104 is coaxially fixedly connected to the rotating block 102 .
[0069] By starting the first motor 104 to rotate the output shaft, the rotating block 102 is driven to rotate, and the rotating block 102 drives the connecting track 11 to rotate, so that the connecting track 11 can be rotated to the other side to dock with the positioning track 1 on the other side, and the connecting surface between the positioning track 1 and the connecting track 11 is an arc-shaped structure, and both are on a circular contour line with the center of the middle positioning track 1 as the center of the circle. Through the setting of the connecting track 11, positioning tracks in different directions can be docked, so that the camera can be driven to move in different directions, so that the same person can be tracked and located, thereby improving the positioning effect.
[0070] Reference Figure 6 and Figure 7 , the embodiment of the present application provides a building intelligent positioning system, the moving component 5 includes a moving block 51, and the cross-section of the moving block 51 is the same as the positioning track 1;
[0071] A plurality of rollers 52 are rotatably arranged outside the moving block 51, one of the moving blocks 51 is arranged in the middle of the moving block 51, and the remaining rollers 52 are arranged in a symmetrical structure;
[0072] A connecting block 53 is fixedly provided at one end of the moving block 51, and the connecting block 53 is connected and fixed to the positioning camera 4;
[0073] A limit block 54 is fixedly disposed on the outer wall of the moving block 51. The limit block 54 is arranged in an annular structure, and the limit block 54 is rotatably connected to a limit groove 55 provided inside the roller 52.
[0074] A gear groove 56 is formed inside the limiting block 54 . A transmission gear 57 is rotatably disposed inside the gear groove 56 . The transmission gear 57 is meshed and connected with an internal gear 58 fixedly disposed inside the limiting groove 55 .
[0075] By setting up the mobile components, the camera can be actively driven to move and can turn in the vertical direction.
[0076] Reference Figure 8 The embodiment of the present application provides a building intelligent positioning system, wherein a mechanical slot 501 is provided inside the moving block 51, and a rotating shaft 503 is rotatably provided inside the mechanical slot 501, and both ends of the rotating shaft 503 are coaxially fixedly connected with upper and lower transmission gears 57 respectively;
[0077] An external gear 504 is disposed on the outer wall of the rotating shaft 503 , and a driving gear 505 is meshedly connected to the outer wall of the external gear 504 . A second motor 506 is coaxially fixedly disposed on one side of the driving gear 505 .
[0078] By starting the second motor 506, the output shaft rotates and drives the driving gear 505 to rotate, and the driving gear 505 is meshed with the external gear 504. At this time, the rotation of the rotating shaft 503 will drive the upper and lower transmission gears 57 to rotate, and the transmission gear 57 is meshed and connected with the internal gear 58 inside the roller 52. At this time, the upper and lower rollers 52 act as active wheels to rotate, thereby driving the moving block 51 and the positioning camera 4 to move and position. This arrangement reduces the occupied space, thereby driving the two transmission gears 57 to rotate.
[0079] When the building intelligent positioning system is needed, first, the facial expression and action data are collected by positioning the camera 4, each image or video frame is manually checked, and then a bounding box or mark is manually drawn through a desktop graphics editing software; computer vision tools and algorithms are used to automatically detect faces and provide preliminary bounding boxes, and then these preliminary marks are manually checked and corrected to ensure accuracy, and key points of the face are marked, such as the positions of eyes, nose, mouth, etc., are marked. At this time, image processing techniques are used to improve image quality, such as denoising, sharpening or brightness adjustment, and the image size is adjusted to have the same size so that the model can handle fixed-size inputs. The diversity of the data is increased by applying random transformations such as rotation, flipping, translation, and scaling to improve the generalization ability of the model. The image data is normalized and the pixel values are scaled to a specific range, usually [0,1] or [-1,1]. During training, additional data enhancement techniques such as random cropping, color enhancement, etc. can be applied to increase the diversity of the data. A suitable model is established using face and action data. Divide into training set, validation set and test set, most of the data is used for training, part is used to verify the performance of the model, and part is used for final testing; when tuning, select an appropriate loss function, which will measure the performance of the model and the accuracy of the prediction, select the stochastic gradient descent SGD optimization algorithm, perform forward propagation and back propagation on the training data to update the weights and parameters of the model, and regularly evaluate the model performance on the validation set to detect overfitting or underfitting; perform hyperparameter adjustment, the hyperparameters include learning rate, batch size, number of layers, number of neurons, regularization strength, etc., use grid search, random search or experience-based methods to adjust the hyperparameters, save and apply the trained tracking and positioning model, at this time, the controller starts the output shaft of the first motor 104 to rotate, driving the rotating block 102 to rotate, and the rotating block 102 will drive the connecting track 11 to rotate, so that the connecting track 11 can rotate to the other side and dock with the positioning track 1 on the other side, and the connecting surface between the positioning track 1 and the connecting track 11 is an arc structure, and both are on the circular contour line with the center of the positioning track 1 in the middle as the center. By starting the second motor 506, the output shaft rotates and drives the driving gear 505 to rotate, and the driving gear 505 is meshed with the external gear 504. At this time, the rotation of the rotating shaft 503 will drive the upper and lower transmission gears 57 to rotate, and the transmission gear 57 is meshed and connected with the internal gear 58 inside the roller 52. At this time, the upper and lower rollers 52 act as active wheels to rotate, thereby driving the moving block 51 and the positioning camera 4 to move for tracking and multiple positioning.
Claims
1. A building intelligent positioning system, characterized in that: Include: The face and motion data collection module collects facial expression and motion data through positioning cameras; The face and motion data marking module marks suspicious facial expressions and motion data; The face and motion data preprocessing module processes the face and motion data to improve the quality of the face and motion data in order to establish a model; Establish a tracking and positioning model, and use face and motion data to build a suitable model; Tracking and positioning model training and tuning module, which uses labeled data sets to train the model and iteratively trains the model, adjusting hyperparameters to improve performance; Save the tracking and positioning model and apply the module to save and apply the trained tracking and positioning model; The mobile execution module can drive the positioning camera for further positioning.
2. The building intelligent positioning system according to claim 1, characterized in that: The face and motion data labeling module uses manual inspection of each image or video frame and then manually draws bounding boxes or labels using desktop graphics editing software; Use computer vision tools and algorithms to automatically detect faces and provide preliminary bounding boxes. These preliminary markings are then manually reviewed and corrected to ensure accuracy. Use the key points of the face to mark, such as the positions of the eyes, nose, mouth, etc.
3. The building intelligent positioning system according to claim 1, characterized in that: The face and action data preprocessing module uses image processing techniques to improve image quality, such as denoising, sharpening or brightness adjustment, adjusts the image size to make it have the same size so that the model can process fixed-size inputs, increases the diversity of data by applying random transformations, improves the generalization ability of the model, normalizes the image data, and scales the pixel values to a specific range, usually [0, 1] or [-1, 1]. During training, additional data enhancement techniques such as random cropping, color enhancement, etc. can be applied to increase the diversity of data.
4. The building intelligent positioning system according to claim 1, characterized in that: The tracking and positioning model training and tuning module divides the data into a training set, a validation set, and a test set, with most of the data used for training, a portion used to verify the performance of the model, and a portion used for final testing; when performing tuning, an appropriate loss function is selected, which will measure the performance of the model and the accuracy of the prediction, and a stochastic gradient descent optimization algorithm is selected to perform forward propagation and back propagation on the training data to update the weights and parameters of the model, and the model performance is regularly evaluated on the validation set to detect overfitting or underfitting; hyperparameter adjustment is performed, and the hyperparameters include learning rate, batch size, number of layers, number of neurons, regularization strength, etc., and grid search, random search, or experience-based methods are used to adjust the hyperparameters.
5. The building intelligent positioning system according to claim 1, characterized in that: The mobile execution module includes Positioning tracks, which can be set at entrances and corridors in different directions; Positioning camera, which can obtain and locate human images; Track connection: positioning tracks in different directions can be connected by track connection; The mobile component can drive the positioning camera to follow the moving position, so that the image can be positioned multiple times continuously.
