A stress sensor arrangement method based on the behavior trajectory of the elderly
By employing a pressure sensor placement method based on the behavioral trajectories of the elderly, and utilizing millimeter-wave radar and image processing technology, the sensor placement is identified and optimized. This solves the problems of monitoring blind spots and equipment waste in traditional placement methods, and achieves efficient and personalized home safety monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ANHUI HEALTH & ELDERLY CARE IND INVESTMENT CO LTD
- Filing Date
- 2026-04-30
- Publication Date
- 2026-07-28
AI Technical Summary
In existing home monitoring systems, the placement of pressure sensors lacks scientific basis, resulting in significant monitoring blind spots and serious equipment waste, failing to meet the needs of modern smart elderly care.
A pressure sensor placement method based on the behavioral trajectories of the elderly was adopted. Activity point clouds were collected by a millimeter-wave radar detector, and the effective activity areas were identified by cubic spline interpolation and ImageJ software. The sensor placement positions were determined by an automated grid numbering algorithm.
It enables precise sensor deployment, reduces the number of devices, avoids monitoring blind spots, provides more accurate and personalized home safety monitoring, and reduces system costs.
Smart Images
Figure CN122469342A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of pressure sensor placement technology, and in particular to a pressure sensor placement method based on the behavioral trajectories of elderly people. Background Technology
[0002] Safety monitoring, behavior recognition, and health early warning for the elderly in their home environments are receiving increasing attention from society and families. Traditional home monitoring methods mainly rely on sensing devices such as cameras, infrared detectors, and wristbands, but these devices have significant shortcomings in terms of privacy, accuracy, coverage, and individual adaptability. Furthermore, to identify the activities of the elderly, many systems use pressure sensors placed at fixed points or evenly distributed over large areas. However, these solutions often lack scientific basis for placement based on real behavioral patterns, frequently resulting in significant blind spots, substantial equipment waste, or the inability to identify key areas, failing to meet the needs of modern smart elderly care. Therefore, deploying pressure sensors based on the behavioral trajectories of the elderly is essential. Summary of the Invention
[0003] To overcome the shortcomings of the prior art, this invention provides a pressure sensor deployment method based on the behavioral trajectories of the elderly, which can accurately identify the activity paths and high-frequency use areas of the elderly at home, thereby achieving the most effective sensor coverage at the most optimized cost and providing the elderly with a more accurate and personalized home safety monitoring solution.
[0004] To achieve the above objectives, the present invention adopts the following technical solution, including: A method for deploying pressure sensors based on the behavioral trajectories of the elderly includes the following steps: S1, Divide the floor plan of the area to be arranged into a grid and draw the grid diagram; S2, In the area to be deployed, use a millimeter-wave radar detector to collect the activity point cloud of the elderly; S3 connects the collected activity point clouds of the elderly into a curve to obtain the behavioral trajectory curve of the elderly; S4, based on the behavioral trajectory curves of the elderly, obtain the activity area of the elderly; S5, based on the activity areas of the elderly, distinguishes between effective activity areas and ineffective activity areas; S6, calculate the number of pressure sensors in each grid; S7, based on the effective activity area of the elderly, extracts the grid where pressure sensors need to be placed; S8. Based on the total number of pressure sensors in each grid calculated in step S6 and the grids where pressure sensors need to be placed extracted in step S7, place the pressure sensors.
[0005] Preferably, in step S2, a millimeter-wave radar detector is used to collect three-dimensional point cloud data of the elderly, and continuous dynamic point cloud is obtained by stitching together multiple consecutive frames.
[0006] Preferably, in step S3, the Z-axis coordinates of the active point cloud are mapped to a two-dimensional plane to obtain a set of two-dimensional discrete data points. Based on the principle of cubic spline interpolation, the discrete data points are fitted into a smooth and continuous behavior trajectory curve using WPS tables.
[0007] Preferably, in step S4, based on the standard shoulder width value for the elderly, the behavior trajectory curve is widened proportionally to both sides to obtain the activity area covered by the trajectory.
[0008] Preferably, in step S5, the active area generated by the behavior trajectory curve is filled with color, and the preset transparency of the active area is adjusted and superimposed to form a color depth difference image. The darker the color, the more effective the active area. The color depth difference image is imported into ImageJ software, converted into a grayscale image, and the total number of pixels and brightness distribution histogram of the whole image are statistically analyzed. The target brightness retention interval is defined by threshold screening, and the effective active area is extracted.
[0009] Preferably, in step S6, the number of pressure sensors in each grid is determined based on the resolution of the number of pressure sensors and the grid size.
[0010] This application also provides a readable storage medium having a computer program stored thereon, which, when executed, implements the aforementioned method for arranging pressure sensors based on the behavioral trajectories of the elderly.
[0011] This application also provides an electronic device including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned method for arranging pressure sensors based on the behavioral trajectories of the elderly.
[0012] This application also provides a computer program product comprising a computer program / instructions that, when executed by a processor, implement the aforementioned method for arranging pressure sensors based on the behavioral trajectories of the elderly.
[0013] The advantages of this invention are: (1) This invention aims to overcome the shortcomings of existing home pressure sensor placement methods for the elderly, such as "blind placement, high cost, inaccurate monitoring, and poor adaptability to individual differences." By introducing millimeter-wave radar point cloud technology, behavioral trajectory modeling technology, image processing technology, and automated grid numbering algorithm, a pressure sensor placement method based on the actual behavioral trajectories of the elderly is proposed. This method can accurately identify the activity paths and high-frequency usage areas of the elderly at home, thereby achieving the most effective sensor coverage at the optimal cost and providing the elderly with a more accurate and personalized home safety monitoring solution.
[0014] (2) This invention is the first to introduce millimeter-wave radar point cloud and trajectory modeling technology into the deployment of pressure sensors. By analyzing the real behavioral trajectories of the elderly, high-frequency activity areas and key passageways are identified, enabling precise deployment of sensors. This method is more scientific and efficient than the traditional uniform deployment method, significantly reducing the number of sensors used while avoiding missed detections.
[0015] (3) This invention obtains high and low activity areas through methods such as trajectory extension, transparency overlay, and brightness analysis, and identifies dark areas as high activity areas, thereby realizing a quantitative expression of activity space. This area classification can directly guide the layout strategy and solves the problem of the lack of a definition of "effective activity area" in traditional methods.
[0016] (4) The present invention utilizes the grayscale distribution and thresholding processing of ImageJ software to filter out high-activity areas, retain effective active areas and remove unimportant areas, thereby making the layout more accurate and more data-driven.
[0017] (5) The present invention uses a grid diagram of fixed size as a spatial basic unit, so that each grid corresponds to a sensor one by one, which facilitates area conversion, path mapping and point recording.
[0018] (6) This invention introduces an image coverage recognition algorithm, which automatically identifies the squares covered by the effective activity area through code program and outputs the specific number of the sensor to be installed, thereby automating the layout decision and reducing human judgment error.
[0019] (7) The present invention uses millimeter-wave radar to collect the trajectory, eliminating the risk of privacy leakage from cameras; at the same time, it reduces the number of sensors, thereby reducing system costs while meeting the requirements of high-resolution monitoring, making it more suitable for home environments.
[0020] (8) After deploying pressure sensors based on the method of this invention, the home activity status of the elderly can be monitored in real time, continuously and without physical contact. By detecting changes in ground force, the passage, stay and duration of the elderly can be determined, and abnormal states such as falls or prolonged stillness can be identified and warned. At the same time, this invention selects high-frequency activity areas based on the actual activity trajectory of the elderly, and deploys pressure sensors only in the effective activity areas, avoiding equipment redundancy and monitoring blind spots caused by traditional uniform deployment, and achieving low-cost, high-precision and personalized home safety monitoring effects. Attached Figure Description
[0021] Figure 1 This is a flowchart of a method for deploying pressure sensors based on the behavioral trajectories of the elderly.
[0022] Figure 2 This is a schematic diagram of a grid drawn based on a floor plan in this embodiment.
[0023] Figure 3 This is a schematic diagram of the behavioral trajectory curve of the elderly in this embodiment.
[0024] Figure 4 This is a schematic diagram of the activity area (trajectory area) of the elderly generated based on behavioral trajectories in this embodiment.
[0025] Figure 5 This is a color depth difference image formed in the activity area of the elderly in this embodiment.
[0026] Figure 6 This is a schematic diagram of the effective activity area for the elderly in this embodiment.
[0027] Figure 7 This is a schematic diagram of the grid where pressure sensors need to be arranged, extracted in this embodiment. Detailed Implementation
[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0029] First, several technical directions closely related to this invention will be introduced.
[0030] I. Behavioral Trajectory Acquisition Technology
[0031] Behavioral trajectory acquisition is a crucial component of intelligent monitoring systems. Its core objective is to accurately obtain the spatial movement trajectory of elderly individuals over a given period of time without disrupting their daily lives. Existing technologies primarily include: 1. Camera-based behavior trajectory acquisition Using cameras to identify human locations and record their trajectories presents serious privacy issues and is also susceptible to factors such as lighting, occlusion, and furniture arrangement, resulting in limited stability.
[0032] 2. Trajectory acquisition based on wearable sensors
[0033] Examples include smart bracelets and inertial measurement units (IMUs). However, elderly people often forget to wear them or refuse to use them due to discomfort, making them unreliable in long-term monitoring scenarios.
[0034] 3. Behavior extraction based on radio frequency (RF) or WiFi sensing
[0035] The presence and movement of a human body can be determined by changes in wireless signals, but the accuracy is currently low and it is difficult to form a continuous trajectory curve.
[0036] II. Millimeter-wave radar point cloud technology
[0037] Millimeter-wave radar, as a non-contact sensing technology, boasts advantages such as strong penetration, no imaging privacy leaks, and minimal susceptibility to lighting conditions. Millimeter-wave radar emits high-frequency electromagnetic waves and calculates the target's distance, angle, and velocity based on echo differences, ultimately forming a point cloud dataset composed of numerous three-dimensional coordinates. However, existing products are primarily used for fall detection, indoor positioning, and presence sensing, rarely for guiding the deployment planning of other devices, and lack a method to explicitly map point cloud trajectories onto a planar grid for spatial decision-making.
[0038] III. Behavioral Trajectory Region Calculation Techniques and Image Processing Techniques
[0039] To obtain the effective activity area of an elderly person from the trajectory, the trajectory lines need to be widened to form the area. Since the point cloud we obtain is generated from the center point of the human body, the obtained trajectory lines can be widened to both sides according to the elderly person's shoulder width to obtain the area swept by these trajectory lines. For example, if the elderly person's shoulder width is 46 cm, then it is only necessary to widen the trajectory lines to both sides by 23 cm to obtain the area traversed by the elderly person. Existing technology involves: 1. Smooth the trajectory using interpolation or spline curves. Using cubic spline curves commonly used in WPS, discrete point clouds can be connected into smooth trajectories, ensuring the accuracy of region calculation.
[0040] 2. Trajectory visualization based on software analysis
[0041] For example, WPS Spreadsheet software can be used to expand the trajectory at equal intervals to simulate the spatial range occupied by human activity.
[0042] 3. Region extraction and area calculation based on grayscale threshold
[0043] Using software such as ImageJ, the pixel distribution of the effective region is obtained through grayscale conversion, threshold segmentation, etc., so as to calculate the area and filter out low participation regions.
[0044] However, existing technologies do not uniformly link the trajectory area with the subsequent sensor deployment logic.
[0045] IV. Pressure Sensor Placement Technology
[0046] Currently, the main methods for deploying pressure sensors in homes include: 1. Uniformly laid layout Sensors can be laid out at equal intervals across the entire ground to achieve high coverage. However, this method is expensive, involves complex wiring, and may result in data redundancy.
[0047] 2. Layout of experience points
[0048] Pressure sensors are only installed in areas such as bedside and bathroom entrance, resulting in a small number of installations. However, based on assumptions, this approach cannot adapt to individual differences, leading to missed detections and monitoring blind spots.
[0049] In practice, the behaviors of the elderly vary greatly; some spend long periods in the living room, while others prefer to linger in the kitchen or on the balcony. Traditional layout methods fail to reflect these differences. There is a lack of technical methods that can accurately arrange layouts based on the actual behavioral patterns of the elderly.
[0050] Next, the technical solution of the present invention will be introduced.
[0051] Depend on Figure 1 As shown, a method for deploying pressure sensors based on the behavioral trajectories of the elderly includes the following steps: S1, Divide the plan of the area to be arranged into a grid and draw the grid diagram.
[0052] In this embodiment, the area to be arranged is the home environment of the elderly, the floor plan is a house floor plan, and the grid is a square grid.
[0053] like Figure 2 The diagram shown is a grid diagram drawn from the floor plan of this embodiment. Each grid cell (square) in the grid diagram is defined as having an area of 0.25 × 0.25 m². 2 .
[0054] S2 uses a millimeter-wave radar detector to collect point clouds of activity data of elderly people.
[0055] The working principle of the millimeter-wave radar detector is as follows, and the entire working process is divided into four parts: (1) First, the signal is generated and transmitted. The frequency synthesizer inside the radar generates millimeter waves of a specific frequency (usually 30GHz-300GHz), and the frequency is finely adjusted according to a preset rule.
[0056] (2) Next is target reflection and echo reception. When millimeter waves encounter a target (such as a pedestrian or obstacle), some of the energy will be reflected to form an echo signal. The echo signal is captured by the radar's receiving antenna array, amplified by a low-noise amplifier (to avoid signal attenuation and loss), and then transmitted to the signal processing unit.
[0057] (3) Next is the calculation of core parameters, which is the core step of radar target perception. All calculations are based on the difference between the echo signal and the transmitted signal, including the calculation of core parameters such as the target's distance, speed, and angle.
[0058] By comparing the time difference between the transmitted signal and the echo signal, and combining this with the speed of electromagnetic waves (speed of light), the target distance is calculated using the formula "d=(Δt×c) / 2", where d is the target distance, Δt is the time difference between the transmitted signal and the echo signal, and c is the speed of light.
[0059] When a millimeter wave is emitted onto a moving object (such as an elderly person walking) and reflected back, the frequency (or phase) of the echo signal undergoes a slight change. If the elderly person is facing the radar, the echo frequency increases; if the elderly person is walking away from the radar, the echo frequency decreases; if the elderly person is stationary, the frequency remains unchanged (velocity is 0). The target velocity can be calculated using the formula "v=(Δf×c) / 2f0", where v is the target velocity and Δf is the frequency difference, Δf=f0-f r f0 is the transmission frequency, f r This is the echo frequency.
[0060] Millimeter-wave radar detectors have a receiving antenna array. If an object (such as an elderly person walking) is directly in front, the echo signal arrives at all receiving antennas simultaneously with a phase difference of 0. If the object is to the side, the echo signal arrives at different receiving antennas at extremely small intervals, resulting in a phase difference. The target angle is calculated using the formula "θ=arcsin(λΔΦ / 2πr)", where θ is the target angle (azimuth), λ is the wavelength, ΔΦ is the phase difference between the two receiving antennas, and r is the distance between the two receiving antennas.
[0061] (4) Finally, there is the output and application of data. The distance, speed and angle obtained through calculation constitute the core information of the target. This core information will be output in the form of digital signals and will eventually be presented in the form of point clouds.
[0062] The image output by a millimeter-wave radar detector is a point cloud map, not a camera image. It is a dataset composed of a large number of three-dimensional coordinate points, with each point corresponding to a reflective unit within the detection area. These points not only contain three-dimensional position information such as X (horizontal distance), Y (lateral offset), and Z (vertical height), but also include attributes such as signal strength (reflectivity) and velocity.
[0063] The mapping principle of the millimeter-wave radar detector is as follows: the generation of active point clouds consists of three steps: (1) First is signal filtering and target extraction. In the echo received by the radar, in addition to the target signal (the reflected echo signal of the elderly), there is also noise (such as electromagnetic interference) and clutter (such as the reflection of the ground and trees). These noise clutters are removed by filtering algorithms (such as Kalman filtering) and only the signal of the effective target is retained.
[0064] (2) Next is the calculation of three-dimensional coordinates. Based on the distance and azimuth calculated earlier, the polar coordinates (distance and angle) are transformed into three-dimensional coordinates of the three-dimensional coordinate system (X, Y, Z).
[0065] (3) Finally, the point cloud is stitched and updated. The radar will continuously transmit signals, and each detection will generate a batch of target points (active point cloud). The target points of multiple consecutive frames are stitched together in time order to form a dynamically updated active point cloud map, which reflects the positional changes of the target (elderly person) in the detection area in real time (e.g., when a person moves in the room, the corresponding point will move in the point cloud map over time).
[0066] In this invention, a millimeter-wave radar detector is placed in the home of an elderly person. The working principle and mapping principle of the millimeter-wave radar detector are used to generate an activity point cloud map of the elderly person's home (the millimeter-wave radar outputs a set of point clouds every tens of milliseconds. After signal processing, the center point position of the human body is calculated for each frame).
[0067] S3 connects the collected activity point clouds of the elderly into a curve to obtain the behavioral trajectory curve of the elderly.
[0068] Mapping the activity point cloud of elderly individuals onto a two-dimensional plane and removing the Z-axis coordinates yields a set of two-dimensional discrete data points. The basic principle of smoothing lines and data marker scatter plots in WPS Spreadsheets is as follows: WPS Spreadsheet uses a cubic spline interpolation method for smoothing lines. The basic idea is to smoothly connect any two known points using a cubic polynomial, making the entire curve continuous and its first and second derivatives continuous. The specific operation is as follows: S31, Assume the known data points are:
[0069] S32, the i-th interval formed between two adjacent data points. In this context, a cubic polynomial is defined as:
[0070] in, , The x-coordinate is the value of the y-coordinate at position x. , , , The coefficients are the coefficients of the cubic polynomial; the coefficients of each interval are obtained by solving a set of constraint equations that satisfy the continuity of function values, the continuity of the first derivative, the continuity of the second derivative, and boundary conditions.
[0071] S33, In order to ensure a smooth connection between all intervals, the constraints of the spline function must be satisfied, including: (1) Each data point:
[0072] (2) The first derivative is continuous:
[0073] (3) The second derivative is continuous:
[0074] (4) Boundary conditions:
[0075] S34, calculate each for all intervals. There is a typical tridiagonal system of equations:
[0076] in,
[0077] Solve all Afterwards, we can obtain:
[0078] Therefore, the curve for each interval can be written as:
[0079] S35, based on the above principles and formulas, WPS Spreadsheet internally draws the data for each interval. Interpolation is calculated using multiple intermediate points, making the image appear as a smooth curve.
[0080] In this invention, the activity points obtained in step S2 are connected in a WPS spreadsheet according to the above principle. The resulting curve represents the activity trajectory of the elderly person at home throughout the day. This trajectory curve appears as many lines in a grid diagram. Curves with dense clusters or high overlap represent routes frequently traversed by the elderly person, while sparser curves or low overlap represent routes with lower activity levels throughout the day. Figure 3 As shown.
[0081] S4, based on the behavioral trajectories of the elderly, determines their activity areas.
[0082] Since the point cloud obtained is generated from the center point of the human body, the obtained trajectory lines can be widened to both sides based on the shoulder width of the elderly person to obtain the area swept by these trajectory lines. For example, if the elderly person's shoulder width is 46 cm, then simply widening the trajectory lines to both sides by 23 cm will give the area traversed by the elderly person. Figure 4 As shown. In real-world detection, there are many of these trajectory lines, and they overlap and intersect extensively, resulting in overlapping surfaces.
[0083] WPS Spreadsheet allows you to widen lines simply by adjusting the line width in points according to a certain ratio.
[0084] S5 distinguishes between effective and ineffective activity areas based on the activity area (trajectory area) of the elderly.
[0085] The activity areas generated by the behavioral trajectory curves will overlap in the graph, creating overlapping areas. Areas with more overlapping areas are the high-activity areas of the elderly. To visually distinguish these high-activity areas, different shades of color are used. Figure 5 As shown, the activity areas generated by each behavioral trajectory curve are colored in light blue. Due to the intricate and mixed trajectory curves in the grid, activity areas of different shades will be generated in the grid diagram. Dark blue and relatively dark blue activity areas are the high activity areas of the elderly, while light blue activity areas are the low activity areas of the elderly. High activity areas are the effective activity areas of the elderly, while low activity areas and inactive areas are considered ineffective activity areas.
[0086] By using WPS Spreadsheet to display the color shades, different active areas can be obtained by changing the transparency of the widened lines from step S4. For example, the line transparency can be adjusted to 50%.
[0087] Extracting the effective activity area of elderly people using ImageJ software: Import the trajectory area generated by the above trajectory lines into ImageJ software. In ImageJ, delete the low-activity areas to obtain the effective activity area. The principle is to directly delete the areas with high brightness coefficients and keep the areas with low brightness coefficients. The specific steps are as follows: (1) Open the image in ImageJ software. The image format is PNG or JPG.
[0088] (2) Change the grayscale value of the image in the software to turn the image into black and white, and use 0 to 255 to represent the brightness (click menu: Image→Type→8-bit), and view the total number of pixels in the image (click menu: Analyze→Measure).
[0089] (3) View the brightness distribution of the whole image (click the menu: Analyze→Histogram). A brightness histogram will appear, which can help filter out areas with excessive brightness.
[0090] (4) Select the brightness range to be retained (click the menu: Image→Adjust→Threshold), adjust the required brightness range (e.g., 80-150), click Apply to apply, and then delete the unwanted brightness area (click the menu: Edit→Selection→Create Selection→Edit→Clear Outside).
[0091] (5) View the retained brightness area and calculate the brightness percentage of the area (click menu: Analyze→Measure). The Area displays the number of pixels in the retained area (unit: pixels). The percentage of the area = number of pixels in the retained area ÷ number of pixels in the whole image × 100%.
[0092] (6) Export the processed image, such as Figure 6 As shown in the image, the area depicted is the effective activity area for the elderly.
[0093] S6. Determine the number of pressure sensors in each grid based on the grid size in the grid diagram.
[0094] Assuming the side length of the grid is L and the required sensor spacing is D, then the number of sensors needed in each grid is N = L. 2 / D 2
[0095] Currently, pressure sensors on the market come in three resolutions: high resolution, which can detect gait and falls; medium resolution, which can detect human activity and presence; and low resolution, which can only detect the presence of a person. The difference in resolution is achieved through the spacing between the sensors; the shorter the distance, the higher the resolution, and vice versa. (High resolution corresponds to a spacing of approximately 0.25m, medium resolution approximately 0.5m, and low resolution approximately 1m.)
[0096] According to the experimental requirements of this embodiment, achieving high resolution is sufficient. To achieve high resolution, the spacing between the pressure sensors needs to be controlled at 0.25m. When high resolution is achieved, the daily activities of the elderly can be detected, thus meeting the requirements of this pressure sensor.
[0097] The spacing between each sensor is currently specified to be 0.25m, since the size of each square in the grid diagram is 0.25m × 0.25m. 2 Therefore, it is sufficient to place one pressure sensor in each square.
[0098] S7 extracts the grid where pressure sensors need to be placed, based on the effective activity area of the elderly.
[0099] The effective active area obtained in step S5 is placed into the grid diagram, and a code is introduced. This code can be used to number each square in the grid diagram. The first square is number 1, the second square is number 2, and so on until all the numbers are assigned. After all the numbers are assigned, the code can automatically identify the imported effective active area and export the square numbers covered by the active area. Pressure sensors are then arranged according to these square numbers.
[0100] In this embodiment, the following code is provided: import cv2 import numpy as np # === Parameters === grid_size = 0.25 # Grid size (meters) scene_width = 10 # Actual width of the venue (meters) scene_height = 8 # Actual site height (meters) coverage_threshold = 0.0 # Coverage threshold (0 means any pixel is acceptable) # === Read Image === img = cv2.imread("activity.png") if img is None: raise FileNotFoundError("Unable to read image file activity.png") gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # === Binarization, Automatic Active Region Identification === _, mask = cv2.threshold(gray, 0, 255, cv2.THRESH_OTSU) mask = (mask > 0).astype(np.uint8) H, W = mask.shape # === Calculate the ratio of image pixels to actual meters === px_per_m_x = W / scene_width px_per_m_y = H / scene_height covered_cells = [] cell_number = 1 num_cols = int(scene_width / grid_size) num_rows = int(scene_height / grid_size) for row in range(num_rows): for col in range(num_cols): # Pixel range of the grid (to prevent out-of-bounds access) x1 = int(col grid_size px_per_m_x) x2 = min(int((col + 1) grid_size px_per_m_x), W) y1 = int(row grid_size px_per_m_y) y2 = min(int((row + 1) grid_size px_per_m_y), H) cell = mask[y1:y2, x1:x2] # Calculate coverage if cell.size > 0: coverage_ratio = cell.sum() / cell.size if coverage_ratio > coverage_threshold: covered_cells.append(cell_number) cell_number += 1 print("Grid numbers where pressure sensors need to be installed:") print(covered_cells) print(f"A total of {len(covered_cells)} cells were covered.") # === Visualization Results === result_img = img.copy() for idx in covered_cells: row = (idx - 1) / / num_cols col = (idx - 1) % num_cols x1 = int(col grid_size px_per_m_x) x2 = min(int((col + 1) grid_size px_per_m_x), W) y1 = int(row grid_size px_per_m_y) y2 = min(int((row + 1) grid_size px_per_m_y), H) cv2.rectangle(result_img, (x1, y1), (x2, y2), (0, 255, 0), 2) cv2.imwrite("result.png", result_img) print("The visualization result has been saved to result.png") For example: Input the active region obtained in step S5 into the above code. As long as the imported active region covers the grid, the grid number will be automatically exported, such as... Figure 7 As shown, the output grid number is: [166, 167, 206, 207, 246, 247, 248, 249, 287, 288, 289, 290, 328,329, 330, 369, 370, 371, 410, 411, 412, 450, 451, 452, 491, 492, 531, 532,533, 572, 573, 574, 612, 613, 614, 615, 653, 654, 655, 656, 684, 694, 695,696, 697, 724, 725, 735, 736, 737, 738, 764, 765, 777, 778, 779, 780, 804, 805, 806, 818, 819, 820, 821, 822, 845, 846, 847, 848, 857, 858, 859, 860, 861, 862, 863, 864, 865, 884, 885, 886, 887, 888, 897, 898, 899, 901, 925, 926, 927, 928, 930, 931, 932, 933, 934, 935, 936, 938, 939, 940, 941, 949, 950, 951, 952, 958, 966, 967, 968, 969, 970, 971, 972, 973, 974, 975, 976, 977, 978, 979, 980, 981, 987, 988, 989, 990, 991, 992, 993, 994, 995, 996, 997, 998, 1007, 1008, 1009, 1010, 1011, 1012, 1013, 1014, 1015, 1016, 1017,1018, 1019, 1020, 1021, 1030, 1031, 1032, 1033, 1034, 1035, 1036, 1037, 1076,1077, 1078, 1117).
[0101] S8. Based on the total number of pressure sensors in each grid calculated in step S6 and the grid number extracted in step S7, arrange the pressure sensors.
[0102] This invention solves the problem of blind placement of traditional pressure sensors. Existing solutions often rely on empirical locations or uniform placement, failing to reflect the individual activity habits of the elderly and easily leading to monitoring blind spots or wasted sensors. The main objective of this invention is to transform the placement strategy from experience-based inference to data-driven approaches through behavioral trajectory analysis.
[0103] This invention reduces sensor deployment costs and redundant equipment. By identifying high-activity areas and critical paths, sensors are deployed only in necessary areas, minimizing the number of sensors, maximizing monitoring efficiency, avoiding a large number of ineffective deployments, and improving overall economic efficiency.
[0104] This invention improves the accuracy and real-time performance of home monitoring for the elderly. By combining millimeter-wave radar point cloud and image processing methods, it generates high-precision behavioral trajectories and obtains effective activity areas based on trajectory region analysis, enabling the monitoring system to more accurately identify the activity status of the elderly.
[0105] This invention enables personalized, dynamic, and adaptable layouts that suit different apartment types and lifestyles. Different elderly people have different spatial usage habits. This invention automatically generates layout schemes through trajectory data, which can adapt to various apartment types and diverse behavioral patterns, avoiding the incompatibility problems caused by the "one-size-fits-all" approach of traditional solutions.
[0106] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for arranging pressure sensors based on the behavioral trajectories of the elderly, characterized in that, Includes the following steps: S1, Divide the floor plan of the area to be arranged into a grid and draw the grid diagram; S2, In the area to be deployed, use a millimeter-wave radar detector to collect the activity point cloud of the elderly; S3 connects the collected activity point clouds of the elderly into a curve to obtain the behavioral trajectory curve of the elderly; S4, based on the behavioral trajectory curves of the elderly, obtain the activity area of the elderly; S5, based on the activity areas of the elderly, distinguishes between effective activity areas and ineffective activity areas; S6, calculate the number of pressure sensors in each grid; S7, based on the effective activity area of the elderly, extracts the grid where pressure sensors need to be placed; S8. Based on the total number of pressure sensors in each grid calculated in step S6 and the grids where pressure sensors need to be placed extracted in step S7, place the pressure sensors.
2. The method for arranging pressure sensors based on the behavioral trajectories of the elderly according to claim 1, characterized in that, In step S2, a millimeter-wave radar detector is used to collect three-dimensional point cloud data of the elderly, and continuous dynamic point cloud is obtained by stitching together multiple consecutive frames.
3. The method for arranging pressure sensors based on the behavioral trajectories of the elderly according to claim 1, characterized in that, In step S3, the Z-axis coordinates of the active point cloud are mapped to a two-dimensional plane to obtain a set of two-dimensional discrete data points. Based on the principle of cubic spline interpolation, the discrete data points are fitted into a smooth and continuous behavior trajectory curve using WPS tables.
4. The method for arranging pressure sensors based on the behavioral trajectories of the elderly according to claim 1, characterized in that, In step S4, based on the standard shoulder width value for the elderly, the behavior trajectory curve is widened proportionally to both sides to obtain the activity area covered by the trajectory.
5. The method for arranging pressure sensors based on the behavioral trajectories of the elderly according to claim 1, characterized in that, In step S5, the active area generated by the behavior trajectory curve is filled with color, and the preset transparency of the active area is adjusted to overlay the display, forming a color depth difference image. The darker the color, the more effective the active area. The color depth difference image is imported into ImageJ software, converted into a grayscale image, and the total number of pixels and brightness distribution histogram of the whole image are statistically analyzed. The target brightness retention interval is defined by threshold screening, and the effective active area is extracted.
6. The method for arranging pressure sensors based on the behavioral trajectories of the elderly according to claim 1, characterized in that, In step S6, the number of pressure sensors in each grid is determined based on the resolution of the number of pressure sensors and the grid size.
7. A readable storage medium, characterized in that, It stores a computer program, which, when executed, implements a pressure sensor arrangement method based on the behavioral trajectory of the elderly as described in any one of claims 1 to 6.
8. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the pressure sensor placement method based on the behavioral trajectory of the elderly as described in any one of claims 1 to 6.
9. A computer program product, characterized in that, It includes a computer program / instruction that, when executed by a processor, implements the pressure sensor placement method based on the behavioral trajectories of the elderly as described in any one of claims 1 to 6.