Indoor Personnel Location Method Based on BIM and Video Surveillance
By creating BIM models of indoor activity spaces and cameras, establishing spatial path maps and mapping relationships, and utilizing existing video surveillance data for intelligent analysis, the problems of high hardware costs and low positioning accuracy in existing technologies have been solved, achieving high-precision indoor personnel positioning and trajectory mapping.
Patent Information
- Application Number
- CN202211568725.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-12-08
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2042-12-08
AI Technical Summary
Existing indoor personnel positioning technologies suffer from high hardware costs, low positioning accuracy, and large errors, making it difficult to meet the needs of daily management and emergency management.
By creating BIM models of indoor activity spaces and cameras, establishing spatial path maps, mapping relationships between grid points and pixels, capturing and locating faces, calculating personnel trajectories, and performing intelligent analysis using existing video surveillance data.
Without increasing hardware costs, the positioning accuracy has been improved to less than 1 meter, and errors such as personnel positioning or trajectory passing through walls or fixed furniture have been resolved, thus reducing hardware costs.
Smart Images

Figure CN116017299B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to an indoor personnel positioning method based on BIM and video surveillance. Background Technology
[0002] Indoor occupancy location tracking during the building's occupancy phase plays a crucial role in both daily management and emergency management. For example, in daily operations, it can be used to analyze employee movement patterns, assess the rationality of office seating arrangements, and determine employee attendance.
[0003] Currently, common indoor positioning technologies mainly rely on people carrying positioning cards to locate them indoors. This requires additional hardware, which is costly. Furthermore, this method has problems such as mismatch between the person and the card, and errors such as the person being located on a wall or table.
[0004] The current conventional approach involves retrieving video surveillance footage and manually searching for suspicious individuals, which is inefficient and labor-intensive. Therefore, some studies are using facial recognition algorithms to analyze existing indoor security video surveillance footage and automatically identify key personnel.
[0005] However, in practical applications, this method can only locate people at the camera's installation location, not their actual location, resulting in a certain degree of error. Especially in public buildings, where the camera's video surveillance range is wide, the error may exceed 50 meters, typically around 5 meters, potentially pinpointing the wrong workstation or office. This makes it difficult to meet the needs of close contact tracing and office efficiency analysis. Furthermore, when mapping personnel trajectories, the continuous movement between two cameras through fixed furniture and walls can lead to trajectory errors, further hindering application requirements. Summary of the Invention
[0006] The purpose of this invention is to provide an indoor personnel positioning method based on BIM and video surveillance.
[0007] To address the above problems, this invention provides an indoor personnel positioning method based on BIM and video surveillance, comprising: Step 1: Create a BIM model of the indoor activity space and the camera to obtain the BIM model of the indoor activity space and the camera; Step 2: Based on the BIM model of the indoor activity space and cameras, create a spatial path map; Step 3: Based on the spatial path map, establish the mapping relationship between grid points and pixel points; Step 4: Based on the mapping relationship between grid points and pixels, perform face capture and localization; Step 5: Calculate the trajectory of the person based on the results of face capture and localization.
[0008] Furthermore, in the above method, step 1: perform BIM modeling on the indoor activity space and the camera to obtain the BIM model of the indoor activity space and the camera, including: Step 1.1: Input the number of rooms r i The wall W i ={w ij}, Top surface t i Ground i The geometric information of the object, and the geometric information of the fixed furniture on the ground; Step 1.2: Input the number of rooms r i Information about camera i, including: the coordinates of camera i's installation location c. i =(lx i ly i , lz i ), camera i facing VC i =(vx i ,vy i ,vz i ) and focal length f i Among them, lz i It is the height relative to the room floor.
[0009] Furthermore, in the above method, step 2: Based on the BIM model of the indoor activity space and the camera, a spatial path map is established, including: Step 2.1: For each room, copy the room's floor and move it vertically upwards by one person's height (Hp) to form the face movement plane (F). i ; Step 2.2: Use a set of equally spaced vertical lines d in the X and Y directions to align with the face motion plane F. i Perform gridding and record the coordinates V of each grid point, i.e., the intersection of the X and Y lines. i ={v ij= {vx ij vy ij Hp}}; Step 2.3: Based on v ij Given the plane coordinates and geometric information of the room walls and furniture, calculate v. ij Is it on the wall or above fixed furniture, and will the point v located on the fixed furniture or wall be... ij The area marked as no standing is allowed; the rest are areas where standing is permitted. Step 2.4: Set the grid point coordinates V i All points v in the stand-alone area ij Add spatial path graph G r ={V r ={v ijr},E r ={e ijrEdge set E r It is V r The edge e consisting of two directly connected points is still considered. ijr A set; a point can be connected to at most four points (up, down, left, and right). If the line connecting two points passes through fixed furniture or walls, it is marked as not connected. From E r Delete Furthermore, in the above method, Hp is 1.6m.
[0010] Furthermore, in the above method, d is 0.1 meters according to the positioning accuracy requirement.
[0011] Furthermore, in the above method, step 3: establishing the mapping relationship between grid points and pixel points based on the spatial path map, includes: Step 3.1: Based on the point coordinates v ij= {vx ij vy ij Hp}, camera coordinates c i =(lx i ly i , lz i ), camera facing VC i =(vx i ,vy i ,vz i ) and focal length f i Calculate the pixel coordinates m of the camera image corresponding to each intersection point in the grid. ij= ( mx ij, my ij) ; Step 3.2 Record the coordinates v of point v ij With pixel coordinates m ij The mapping relationship M={(v ij m ij )}.
[0012] Furthermore, in the above method, in step 3.1, by adding c to the BIM model... i Set up a camera at the location and set its orientation to vc. i The focal length is set to f. i Then, the camera's perspective is simulated in the model, and a view is rendered; this view is basically consistent with the monitoring screen in the camera; finally, the pixel coordinates of each grid point in the model are recorded in the view. ( mx ij, my ij) .
[0013] Furthermore, in the above method, step 4: based on the mapping relationship between grid points and pixels, face capture and localization are performed, including: Step 4.1: Enter facial images of all key personnel. k This forms a face database F={f k},like Figure 3 As shown; Step 4.2: Use a face capture server to access the images from each camera in real time, such as... Figure 4 As shown; Step 4.3: Calculate and recognize facial images from each camera in real time based on the face database F, and record f. kq = (personnel f) k Time t kq Camera C kq screen coordinates fm kq ), and add it to the set FT={ft k =( f k ,{f kq})}, where ft k Personnel f were recorded k The trajectory information, and based on time t kq Sort by morning to evening; Step 4.4: For each person f kq Calculate the distance coordinate fm from the mapping relationship M. kq The nearest element (v) kq m kq ); will v kq As f kq The actual location point is added to f. kq =( face f k Time t kq Camera C kq screen coordinates fm kq v kq ).
[0014] Furthermore, in the above method, step 5: based on the results of face capture and localization, calculate the person's trajectory, including: Step 5.1: Iterate through each element ft in FT k ; Step 5.2: Traverse ft k Each element f in kq ; Step 5.3: Based on the spatial path map G r Using the shortest path algorithm Dijkstra's algorithm from f kq grid points v kq At the next moment f kq+1 grid points v kq+1 And record the grid points VP that are passed through. kq ={v kqp}, p=1,2,…,P, where P is the number of cells from v kq to v kq+1 The number of grid points traversed; Step 5.4: Assume personnel f k average duration tp kq = After VP kq Each point v in kqp and VP kq Information f for each point kqp =(face f k Time t kq +p*tp kq Camera C kq , / , v kqp Add to ft k Chinese f kq and f kq+1 between; Step 5.5: Traverse ft k After processing each element, return to step 5.1; Step 5.6: After traversing all elements in the FT, output the FT, including: each person's f k The movement of people is traced by the grid points they pass through, based on the time sequence of their passage through the grid points.
[0015] Compared with existing technologies, this invention obtains BIM models of indoor activity spaces and cameras by performing BIM modeling; based on these BIM models, a spatial path map is established; based on the spatial path map, a mapping relationship between grid points and pixels is established; based on this mapping relationship, face capture and positioning are performed; and based on the results of face capture and positioning, personnel trajectories are calculated. This patent, without increasing additional hardware costs, utilizes a large amount of existing video surveillance data within buildings to achieve image-based person search, personnel positioning, and trajectory drawing through intelligent algorithms, supporting personnel tracking and analysis in emergency situations. Compared with existing methods, it improves personnel positioning accuracy, with an error of less than 1 meter; it solves errors caused by personnel positioning or trajectories passing through walls, fixed furniture, etc.; it makes full use of existing equipment, requiring less additional equipment and reducing the hardware cost of personnel positioning. Attached Figure Description
[0016] Figure 1 This is a schematic diagram of a BIM model according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the monitoring screen of a camera according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a face database according to an embodiment of the present invention; Figure 4This is a schematic diagram of face capture in video surveillance according to an embodiment of the present invention; Figure 5 This is a schematic diagram of personnel trajectory drawing according to an embodiment of the present invention; Figure 6 This is a flowchart of an indoor personnel positioning method based on BIM and video surveillance according to an embodiment of the present invention. Detailed Implementation
[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0018] like Figures 1 to 6 As shown, this invention provides an indoor personnel positioning method based on BIM and video surveillance, comprising: Step 1: Create a BIM model of the indoor activity space and the camera to obtain the BIM model of the indoor activity space and the camera; Step 2: Based on the BIM model of the indoor activity space and cameras, create a spatial path map; Step 3: Based on the spatial path map, establish the mapping relationship between grid points and pixel points; Step 4: Based on the mapping relationship between grid points and pixels, perform face capture and localization; Step 5: Calculate the trajectory of the person based on the results of face capture and localization.
[0019] Preferably, step 1: Perform BIM modeling on the indoor activity space and the camera to obtain the BIM model of the indoor activity space and the camera, including: Step 1.1: Input the number of rooms r i The wall W i ={w ij}, Top surface t i Ground i The geometric information of the object, and the geometric information of the fixed furniture on the ground; Here, as Figure 1 As shown, W1 = {wall w1, wall w2}, top surface t1, floor l1, fixed furniture 1, fixed furniture 2, etc. Step 1.2: Input the number of rooms r i Information about camera i, including: the coordinates of camera i's installation location c. i =(lx i ly i , lz i ), camera i facing VC i =(vx i ,vy i ,vz i ) and focal length fi Among them, lz i It is the height relative to the room floor; Here, as Figure 1 As shown, the coordinates of camera i are c1=(0,0,3.2), and its orientation is vc=(1,1,-0.2).
[0020] Preferably, step 2: Based on the BIM model of the indoor activity space and the camera, establish a spatial path map, including: Step 2.1: For each room, copy the room's floor and move it vertically upwards by one person's height (Hp) to form the face movement plane (F). i ; The preferred value for Hp is 1.6m. Step 2.2: Use a set of equally spaced vertical lines d in the X and Y directions to align with the face motion plane F. i Perform gridding and record the coordinates V of each grid point, i.e., the intersection of the X and Y lines. i ={v ij= {vx ij vy ij Hp}}; The preferred value for d is 0.1 meters; like Figure 1 As shown, the face activity plane F1 is meshed, and the grid point coordinates include V. i ={v 11 v 12 v 13 v 21 v 22 v 23 , ..., v 31 v 32 v 33}; Step 2.3: Based on v ij Given the plane coordinates and geometric information of the room walls and furniture, calculate v. ij Is it on the wall or above fixed furniture, and will the point v located on the fixed furniture or wall be... ij The area marked as no standing is allowed; the rest are areas where standing is permitted. like Figure 1 As shown, v 13 v 23 and v 33 , is a point where people cannot stand; the other points are points where people can stand. Step 2.4: Set the grid point coordinates V i All points v in the stand-alone area ij Add spatial path graph G r ={V r ={v ijr},Er ={e ijr Edge set E r It is V r The edge e consisting of two directly connected points is still considered. ijr A set; a point can be connected to at most four points (up, down, left, and right). If the line connecting two points passes through fixed furniture or walls, it is marked as not connected. From E r Delete; like Figure 1 As shown, G r ={{ v 11 v 12 v 21 v 22 v 31 v 32},{( v 11 v 21 ), ( v 21 v 31 ),( v 11 v 12 ),….}}.
[0021] Preferably, step 3: Based on the spatial path map, establish the mapping relationship between grid points and pixel points, including: Step 3.1: Based on the point coordinates v ij= {vx ij vy ij Hp}, camera coordinates c i =(lx i ly i , lz i ), camera facing VC i =(vx i ,vy i ,vz i ) and focal length f i Calculate the pixel coordinates m of the camera image corresponding to each intersection point in the grid. ij= ( mx ij, my ij) ; Preferably, by using c in the BIM model i Set up a camera at the location and set its orientation to vc. i The focal length is set to f. i Then, simulate the camera's perspective in the model and render a view, such as... Figure 1 As shown.
[0022] Step 3.2 Record the coordinates v of point v ij With pixel coordinates m ij The mapping relationship M={(v ij mij )}, like Figure 2 As shown, M = {(v 11 m 11 ), (v 12 m 12 ), (v 21 m 21 ), (v 22 m 22 ), (v 23 m 23 )…,( v 33 m 33 )}.
[0023] Preferably, step 4: Based on the mapping relationship between grid points and pixels, perform face capture and localization, including: Step 4.1: Enter facial images of all key personnel. k This forms a face database F={f k},like Figure 3 As shown; Step 4.2: Use a face capture server to access the images from each camera in real time, such as... Figure 4 As shown; Step 4.3: Calculate and recognize facial images from each camera in real time based on the face database F, and record f. kq = (personnel f) k Time t kq Camera C kq screen coordinates fm kq ), and add it to the set FT={ft k =( f k ,{f kq})}, where ft k Personnel f were recorded k The trajectory information, and based on time t kq Sort by morning to evening; like Figure 3 As shown, FT={ft1=( f1,{ (f1, 2022-09-26 12:01:52 ,C1,m 11 ), (f1, 2022-09-26 12:02:30, C1, m 31 )})}; Step 4.4: For each person f kq Calculate the distance coordinate fm from the mapping relationship M. kq The nearest element (v) kq m kq ); will v kq As f kqThe actual location point is added to f. kq =( face f k Time t kq Camera C kq screen coordinates fm kq v kq ); like Figure 3 As shown, FT={ft1=(f1,{ft 11 =(f1, 2022-09-26 12:01:52, C1, m 11 v 11 ), ft 12 =(f1, 2022-09-26 12:02:30, C1, m 31 v 31 )}), ft2=( f2,{ f 21= (f2, 2022-09-26 12:01:52, C1, m 11 v 11 ), f 22= (f2, 2022-09-26 12:02:30, C1, m 31 v 31 )})}.
[0024] Preferably, step 5: Based on the results of face capture and localization, calculate the person's trajectory, including: Step 5.1: Iterate through each element ft in FT k ;like Figure 4 As shown, ft2; Step 5.2: Traverse ft k Each element f in kq ;like Figure 4 As shown, f 21, f 22 ; Step 5.3: Based on the spatial path map G r Using the shortest path algorithm Dijkstra's algorithm from f kq grid points v kq At the next moment f kq+1 grid points v kq+1 And record the grid points VP that are passed through. kq ={v kqp}, p=1,2,…,P, where P is the number of cells from v kq to v kq+1 The number of grid points traversed; like Figure 3 As shown, from f 21 v 11 to f 22v 31 It needs to go through v 21 ;VP 21 ={ v 21}; Step 5.4: Assume personnel f k average duration tp kq = After VP kq Each point v in kqp and VP kq Information f for each point kqp =(face f k Time t kq +p*tp kq Camera C kq , / , v kqp Add to ft k Chinese f kq and f kq+1 between; like Figure 3 As shown, tp 21 =19 seconds, VP 21 Add to ft2=( f2,{ f 21= (f2, 2022-09-26 12:01:52, C1, m 11 v 11 ), (f2, 2022-09-26 12:02:11, C1, / , v 21 ), f 22= (f2, 2022-09-26 12:02:30, C1, m 31 v 31 )}); Step 5.5: Traverse ft k After processing each element, return to step 5.1; Step 5.6: After traversing all elements in the FT, output the FT, including: each person's f k The movement of people is traced by the grid points they pass through, based on the time sequence of their passage through the grid points.
[0025] like Figure 5 As shown, personnel f2 are displayed, from v 11 to v 21 Then to v 31 The path.
[0026] In summary, this invention addresses the shortcomings of current location methods based on video surveillance personnel identification by proposing an indoor personnel location method based on Building Information Modeling (BIM) and intelligent video surveillance analysis. This patent utilizes a large amount of existing video surveillance data within buildings, employing intelligent algorithms to achieve image-based personnel search, location, and trajectory mapping, supporting personnel tracking and analysis in emergency situations, without increasing additional hardware costs. Compared to existing methods, it improves personnel location accuracy to less than 1 meter; it eliminates errors caused by personnel tracing paths through walls or fixed furniture; and it fully utilizes existing equipment, requiring minimal additional equipment and reducing hardware costs for personnel location.
[0027] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0028] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0029] Obviously, those skilled in the art can make various modifications and variations to the invention without departing from the spirit and scope of the invention. Therefore, if these modifications and variations fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.
Claims
1. A method for indoor personnel positioning based on BIM and video surveillance, characterized in that, include: Step 1: Create a BIM model of the indoor activity space and the camera to obtain the BIM model of the indoor activity space and the camera; Step 2: Based on the BIM model of the indoor activity space and cameras, create a spatial path map; Step 3: Based on the spatial path map, establish a mapping relationship between spatial grid points and pixels in the camera monitoring image; Step 4: Based on the mapping relationship between grid point coordinates and pixel points, perform face capture and localization; Step 5: Calculate the person's trajectory based on the face capture and localization results; Step 1: Create a BIM model of the indoor activity space and the camera, resulting in a BIM model of the indoor activity space and the camera, including: Step 1.1: Input the number of rooms r i The wall W i ={w ij }, Top surface t i Ground i The geometric information of the object, and the geometric information of the fixed furniture on the ground; Step 1.2: Input the number of rooms r i Information about camera i', including: the coordinates of camera i''s installation location c. i’ =(lx i’ ly i’ lz i’ ), camera i' facing vc i’ =(vx i’ ,vy i’ ,vz i’ ) and focal length f i’ Among them, lz i’ It is the height relative to the room floor; Step 2: Based on the BIM model of the indoor activity space and cameras, create a spatial path map, including: Step 2.1: For each room, copy the room's floor and move it vertically upwards by one person's height (Hp) to form the face movement plane (F). i ; Step 2.2: Use a set of equally spaced vertical lines d in the X and Y directions to align with the face motion plane F. i Perform gridding and record the coordinates V of each grid point, i.e., the intersection of the X and Y lines. i ={v ij= {vx ij vy ij Hp}}; Step 2.3: Based on v ij Given the plane coordinates and geometric information of the room walls and furniture, calculate v. ij Is it on the wall or above fixed furniture, and will the point v located on the fixed furniture or wall be... ij The area marked as no standing is allowed; the rest are areas where standing is permitted. Step 2.4: Set the grid point coordinates V i All points v in the stand-alone area ij Add spatial path graph G r ={V r ={v ijr },E r ={e ijr Edge set E r It is V r An edge e consisting of any two directly connected points in the equation ijr A set; a point can be connected to at most four points (up, down, left, and right). If the line connecting two points passes through fixed furniture or walls, it is marked as not connected. From E r Delete it.
2. The indoor personnel positioning method based on BIM and video surveillance as described in claim 1, characterized in that, Hp is 1.6m.
3. The indoor personnel positioning method based on BIM and video surveillance as described in claim 1, characterized in that, Based on the indoor positioning accuracy requirements, d is 0.1 meters.
4. The indoor personnel positioning method based on BIM and video surveillance as described in claim 1, characterized in that, Step 3: Based on the spatial path map, establish the mapping relationship between grid points and pixel points, including: Step 3.1: Based on the point coordinates v ij= {vx ij vy ij Hp}, camera coordinates c i’ =(lx i’ ly i’ lz i’ ), camera facing VC i’ =(vx i’ ,vy i’ ,vz i’ ) and focal length f i Calculate the pixel coordinates m of the camera image corresponding to each intersection point in the grid. ij’= ( mx ij’, my ij’) ; Step 3.2 Record the coordinates v of point v ij With pixel coordinates m ij’ The mapping relationship M={(v ij m ij’ )}.
5. The indoor personnel positioning method based on BIM and video surveillance as described in claim 1, characterized in that, In step 3.1, by adding c to the BIM model i’ Set up a camera at the location and set its orientation to vc. i’ The focal length is set to f. i’ Then, in BIM, the camera's perspective is simulated to render a two-dimensional view; this view is basically consistent with the monitoring screen in the camera. Finally, record the pixel coordinates of each grid point in the model within the view. ( mx ij, my ij) .
6. The indoor personnel positioning method based on BIM and video surveillance as described in claim 4, characterized in that, Step 4: Based on the mapping relationship between grid points and pixels, perform face capture and localization, including: Step 4.1: Enter facial images of all key personnel. k This forms a face database F={f k }; Step 4.2: Use a face capture server to access the images from each camera in real time; Step 4.3: Calculate and recognize facial images from each camera in real time based on the face database F, and record f. kq = (personnel f) k Time t kq Camera C kq screen coordinates fm kq ), and add it to the set FT={ft k =( f k ,{f kq })}, where ft k Personnel f were recorded k The trajectory information, and based on time t kq Sort by morning to evening; Step 4.4: For each person f kq Calculate the distance coordinate fm from the mapping relationship M. kq The nearest element (v) kq m kq ); will v kq As f kq The actual location point is added to f. kq =( face f k Time t kq Camera C kq screen coordinates fm kq v kq ).
7. The indoor personnel positioning method based on BIM and video surveillance as described in claim 4, characterized in that, Step 5: Based on the results of face capture and localization, calculate the person's trajectory, including: Step 5.1: Iterate through each element ft in FT k ; Step 5.2: Traverse ft k Each element f in kq ; Step 5.3: Based on the spatial path map G r Using the shortest path algorithm Dijkstra's algorithm from f kq grid points v kq At the next moment f kq+1 grid points v kq+1 And record the grid points VP that are passed through. kq ={v kqp }, p=1,2,…,P, where P is the number of cells from v kq to v kq+1 The number of grid points traversed; Step 5.4: Assume personnel f k average duration tp kq = After VP kq Each point v in kqp and VP kq Information f for each point kqp =(face f k Time t kq +p*tp kq Camera C kq v kqp Join ft k Chinese f kq and f kq+1 between; Step 5.5: Traverse ft k After processing each element, return to step 5.1; Step 5.6: After traversing all elements in the FT, output the FT, including: each person's f k The movement of people is traced by the grid points they pass through, based on the time sequence of their passage through the grid points.
Citation Information
Patent Citations
A personnel track tracing system and method based on BIM and GIS video monitoring
CN109886196A
Method for dynamically tracking and positioning indoor personnel in large-scale place
CN111080679A
