Non-image acquisition monitoring system for dangerous actions of old people

Abstract

The invention belongs to the technical field of elderly monitoring, and discloses an elderly dangerous action non-image acquisition monitoring system, an industrial lens is arranged at the front end of a shell, a single-pixel detector is arranged in the shell, and a code modulation disc is also arranged in the shell and located between the industrial lens and the single-pixel detector; a matrix pattern arranged according to a set rule is arranged on the code modulation disc, a servo motor driving the code modulation disc to rotate is arranged in the shell, and the industrial lens gathers collected two-dimensional images to a modulation pattern center on the code modulation disc to modulate light intensity information of the two-dimensional images. And a single-pixel detector behind the coding modulation disc is used for receiving modulation light intensity information of the coding modulation disc. According to the method, dangerous behaviors and actions of the old people are deduced through a computational imaging method, the image reconstruction process is omitted, and then the privacy problem caused by image acquisition is avoided.

Description

Technical Field

[0001] This invention relates to the field of elderly monitoring technology, and in particular to a video-free monitoring system for dangerous actions of the elderly. Background Technology

[0002] With the increasing aging population in China, the demand for home care among the elderly is growing. The purpose of developing an elderly abnormal behavior monitoring system is to monitor the elderly's movement and vital signs in real time, and to issue alarm signals in case of discomfort or accidents, promptly notifying children, hospitals, communities, relatives, neighbors, etc., to provide assistance, thus achieving intelligent care for the elderly. Current monitoring technologies rely on the acquisition of images and videos, analyzing the elderly's actual movements through real-time visuals to determine if dangerous behavior exists and trigger alarms. This type of real-time video acquisition method infringes on the privacy rights of the elderly, leading many elderly people to resist the installation of monitoring equipment. Summary of the Invention

[0003] To overcome the technical defects of existing technologies, this invention provides a monitoring system for dangerous actions of the elderly without image acquisition. It uses computational imaging methods to infer dangerous behaviors of the elderly, omitting the image reconstruction process and thus avoiding privacy issues caused by image acquisition.

[0004] The technical solution adopted in this invention is: a video-free monitoring system for dangerous actions of the elderly, comprising a housing, an industrial lens, an encoding modulation disk, and a single-pixel detector;

[0005] The industrial lens is located at the front end of the housing, the single-pixel detector is located inside the housing, and the encoding modulation disk is also located inside the housing, between the industrial lens and the single-pixel detector.

[0006] The encoding modulation disk has a matrix pattern set according to a predetermined rule, and the inside of the housing is equipped with a servo motor that drives the encoding modulation disk to rotate.

[0007] The industrial lens gathers the acquired two-dimensional image onto the center of the modulation pattern on the encoding modulation disk to modulate the light intensity information of the two-dimensional image, and uses the single-pixel detector behind the encoding modulation disk to receive the modulated light intensity information of the encoding modulation disk.

[0008] During the imaging process, the encoding modulation disk is driven to rotate by a servo motor, forming a regular periodic rotation. Different patterns modulate the target light field in sequence, and the single-pixel detector synchronously collects the modulated light intensity value. The light intensity value is analyzed by deep learning to capture the elderly's movements and identify various dangerous postures.

[0009] Preferably, the response wavelength range of the single-pixel detector is 0.4-1.1µm, and the single-pixel detector is coaxial with the matrix pattern at the stop position of the encoder modulation disk controlled by the servo motor each time it rotates.

[0010] Preferably, the encoding modulation disk is made of a light-shielding material with a diameter of 10 mm and a thickness of 0.5-1 mm, and the matrix pattern is distributed in a polar array near the edge of the encoding modulation disk.

[0011] Preferably, the center of the encoding modulation disk is provided with multiple connection holes for connecting to the output end of the servo motor.

[0012] Preferably, the matrix pattern is a cyclic "S" matrix mask. The "S" matrix mask is generated by cyclically shifting an initial binary encoding sequence of a design. Each time, the sequence is shifted one position to the right, and multiple encoding patterns with cyclic structures are constructed in sequence. These patterns are then evenly distributed on different sectors of the rotating mask to form the "S" matrix mask, and the spacing between the patterns is equal.

[0013] Preferably, the initial binary encoded sequence is generated by a pseudo-random method or a twin prime number algorithm.

[0014] Preferably, the housing also includes an AC-DC regulated power supply module and a control circuit module, and the control circuit module includes a drive circuit that controls the servo motor to drive the encoding modulation disk to rotate periodically. The AC-DC regulated power supply module is used to supply power to the control circuit module and the single-pixel detector.

[0015] The housing has a threaded interface, through which the industrial lens is connected to the housing.

[0016] The servo motor rotates at speeds on the order of centimeters.

[0017] The beneficial effects of this invention are:

[0018] 1. This invention forms the main body of a monitoring system using an industrial lens, an encoding modulation disk, and a single-pixel detector. The encoding modulation disk is also located inside the housing, between the industrial lens and the single-pixel detector. It converges the acquired two-dimensional image onto the center of a modulation pattern on the encoding modulation disk to modulate the light intensity information of the two-dimensional image. The single-pixel detector behind the encoding modulation disk receives this modulated light intensity information and simultaneously collects the modulated light intensity value. Deep learning analysis of these light intensity values ​​is used to capture the elderly person's movements and identify various dangerous postures. By collecting the elderly person's contour information and combining it with corresponding algorithms, even at a low sampling rate, it can effectively determine whether the elderly person has engaged in dangerous behavior. Through one-dimensional light intensity sampling combined with deep learning algorithms, it can accurately identify various postures such as falls, slow movements, and rollovers. Compared to traditional cameras, this technology achieves monitoring through contour feature analysis, avoiding privacy issues associated with image acquisition, making it particularly suitable for private spaces such as bathrooms and bedrooms. Attached Figure Description

[0019] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0020] Figure 1 This is a schematic diagram of the overall structure of the present invention;

[0021] Figure 2 This is a schematic diagram of the internal structure of the housing of the present invention;

[0022] Figure 3 This is a schematic diagram of the structure of the encoding modulation disk of the present invention;

[0023] Figure 4 This is a schematic diagram of the power supply for the present invention;

[0024] Figure 5 This is a schematic diagram illustrating the working principle of the present invention.

[0025] Explanation of reference numerals in the attached figures: 1. Housing; 2. Industrial lens; 3. Encoding modulation disk; 4. Single pixel detector; 5. Matrix pattern; 6. Servo motor; 7. Connection hole; 8. AC-DC regulated power supply module; 9. Control circuit module. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the various embodiments of this invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the various embodiments of this invention to facilitate a better understanding of this application. However, the technical solutions claimed in the claims of this application can be implemented even without these technical details and with various variations and modifications based on the following embodiments.

[0027] like Figures 1 to 5 As shown, this embodiment provides a video-free monitoring system for dangerous actions of the elderly, including a housing 1, an industrial lens 2, an encoding modulation disk 3, and a single-pixel detector 4.

[0028] The industrial lens 2 is disposed at the front end of the housing 1, and the single-pixel detector 4 is disposed inside the housing 1. The encoding modulation disk 3 is also disposed inside the housing 1 and is located between the industrial lens 2 and the single-pixel detector 4.

[0029] The encoding modulation disk 3 is provided with a matrix pattern 5 arranged according to a set rule, and the inside of the housing 1 is provided with a servo motor 6 that drives the encoding modulation disk 3 to rotate regularly and periodically.

[0030] The entire housing 1 has a cubic structure and a 30mm diameter threaded interface for mounting the industrial lens 2. The industrial lens 2 is connected to the housing 1 via the threaded interface. This invention mainly consists of four parts: the housing 1, the industrial lens 2, the encoding modulation disk 3, and the single-pixel detector 4. The encoding modulation disk is mounted on the rear side of the industrial lens and has a connection hole in its center. It is integrated with a servo motor via this connection hole. The servo motor is fixed to the bottom of the housing via a support rod. An AC-DC regulated power supply module and a control circuit module are mounted behind the servo motor, connected by screws to form a single unit.

[0031] The industrial lens 2 focuses the acquired two-dimensional image onto the center of the modulation pattern on the encoding modulation disk 3 to modulate the light intensity information of the two-dimensional image. A single-pixel detector behind the encoding modulation disk receives this modulated light intensity information. The incident light from the image of the target object is focused by the industrial lens and converged onto the surface of the encoding modulation disk. A servo motor controls the rotation of the encoding modulation disk, creating a regular, periodic rotation controlled by a preset drive circuit. The light converged onto the surface of the encoding modulation disk and passes through the rotating disk, forming periodically changing modulated light. This modulated light is received by the single-pixel detector, which collects the light intensity information. Feature extraction can then be performed based on this light intensity information.

[0032] First, an infrared light source illuminates the human body, and a single-pixel detector receives the reflected light. Movement causes fluctuations in light intensity, forming a time-series signal. Motion features are extracted, and time-domain analysis and signal waveform analysis are performed to identify specific waveform combinations: a sudden drop in light intensity (occlusion) + a sustained low signal (falling and remaining still). Spatial domain correlation is then applied. The detection area is divided into an m×n virtual grid, and a modulator sequentially projects coded light onto each grid. Motion localization: hand movement causes a sudden change in signal at a specific grid, with trajectory fitting accuracy reaching ±5cm.

[0033] Deep learning enhancement, model architecture:

[0034] Input: One-dimensional light intensity sequence → 1D-CNN extracts temporal features → LSTM captures action continuity → Fully connected layer classification. Dynamically compare the current signal It with the initial signal I0 to calculate the attenuation rate. If δ>80% and lasts for more than 500ms, a fall alarm will be triggered.

[0035] During the imaging process, the mask rotates with the motor, and different patterns sequentially modulate the target light field. Single-pixel detectors synchronously collect light intensity values, and these cyclically coded patterns are ultimately combined for image reconstruction. To ensure imaging quality, the initial binary coded sequence is typically generated using a pseudo-random twin prime number algorithm. These sequences ensure low correlation and good modulation performance between different patterns, forming the basis for high-quality image reconstruction. To protect the privacy of the elderly, this invention uses image-free sampling technology to calculate and reconstruct their behavioral actions.

[0036] During the imaging process, the coded modulation disk is driven by a servo motor to rotate in a regular, periodic motion. Different patterns sequentially modulate the target light field, and a single-pixel detector synchronously collects the modulated light intensity values. Deep learning analysis of these light intensity values ​​captures the elderly person's movements to identify various dangerous postures. This technology utilizes computational correlation imaging, requiring only one-dimensional light intensity sampling. By calculating and inferring the elderly person's movements, it avoids the image reconstruction process. Specifically, by collecting the elderly person's contour information and combining it with appropriate algorithms, even at low sampling rates, it can effectively determine whether the elderly person has engaged in dangerous behavior. Through one-dimensional light intensity sampling combined with deep learning algorithms, it can accurately identify various postures such as falls, slow movements, and rollovers. Compared to traditional cameras, this technology achieves monitoring through contour feature analysis, avoiding privacy issues associated with image capture, making it particularly suitable for private spaces such as bathrooms and bedrooms.

[0037] The single-pixel detector 4 has a response wavelength range of 0.4-1.1µm. The single-pixel detector 4 is coaxial with the matrix pattern 5 at the stop position of the encoding modulation disk 3, which is controlled by the servo motor 6 each time it rotates. In a single-pixel detection system, the encoding modulation disk typically refers to a mechanical or optical component that spatially modulates the light field through a physical structure. By rotating or translating the encoded pattern (such as a pseudo-random binary matrix or Hadamard code) on the disk surface, the spatial information of the target object is converted into a temporal light intensity signal, thus eliminating the need for image reconstruction.

[0038] The encoding modulation disk 3 is made of a light-shielding material with a diameter of 10 mm and a thickness of 0.5-1 mm, and the matrix pattern 5 is distributed in a polar array near the edge of the encoding modulation disk 3.

[0039] The center of the encoding modulation disk 3 is provided with multiple connection holes 7 for connecting to the output end of the servo motor 6, which facilitates the installation and docking of the encoding modulation disk 3 and the servo motor 6.

[0040] The matrix pattern is a cyclic "S" matrix mask. The "S" matrix mask is generated by cyclically shifting an initial binary encoding sequence. Each time, the sequence is shifted one position to the right, constructing multiple encoding patterns with a cyclic structure. These patterns are then evenly distributed across different sectors of the rotating mask to form the "S" matrix mask, with equal spacing between patterns. The initial binary encoding sequence is generated using a pseudo-random method or a twin prime number algorithm. The housing 1 also houses an AC-DC regulated power supply module 8 and a control circuit module 9. The control circuit module 9 contains a drive circuit that controls the servo motor 6 to periodically rotate the encoding modulation disk 3. The AC-DC regulated power supply module 8 supplies power to the control circuit module 9 and the single-pixel detector 4. The housing 1 has a threaded interface, through which the industrial lens 2 connects to the housing 1. The servo motor 6 rotates at centimeter speeds.

[0041] Furthermore, any content not described in detail in this specification is existing technology known to those skilled in the art.

[0042] Those skilled in the art will understand that the above embodiments are specific examples of implementing the present invention, and in practical applications, various changes in form and detail may be made without departing from the spirit and scope of the present invention.

Claims

1. A video-free monitoring system for dangerous actions of the elderly, characterized in that: It includes a housing (1), an industrial lens (2), an encoding modulation disk (3), and a single-pixel detector (4); The industrial lens (2) is located at the front end of the housing (1), and the single-pixel detector (4) is located inside the housing (1). The encoding modulation disk (3) is also located inside the housing (1) and between the industrial lens (2) and the single-pixel detector (4). The encoding modulation disk (3) is provided with a matrix pattern (5) arranged according to a set rule, and the inside of the housing (1) is provided with a servo motor (6) that drives the encoding modulation disk (3) to rotate regularly and periodically. The industrial lens (2) gathers the acquired two-dimensional image onto the center of the modulation pattern on the encoding modulation disk (3) to modulate the light intensity information of the two-dimensional image, and uses the single-pixel detector behind the encoding modulation disk to receive the modulated light intensity information of the encoding modulation disk. During the imaging process, the encoding modulation disk (3) is driven to rotate by the servo motor (6) and forms a regular periodic rotation. Different patterns modulate the target light field in sequence. The single pixel detector synchronously collects the modulated light intensity value and performs deep learning analysis on the light intensity value to capture the elderly's movements and identify various dangerous postures.

2. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The response wavelength range of the single-pixel detector (4) is 0.4-1.1µm. The single-pixel detector (4) is coaxial with the matrix pattern (5) at the stop position of the encoder modulation disk (3) controlled by the servo motor (6) each time it rotates.

3. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The encoding modulation disk (3) is made of light-shielding material with a diameter of 10 mm and a thickness of 0.5-1 mm. The matrix pattern (5) is located near the edge of the encoding modulation disk (3) and is distributed in a polar array.

4. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The center of the encoding modulation disk (3) is provided with multiple connection holes (7) for connecting to the output end of the servo motor (6).

5. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The matrix pattern is a cyclic "S" matrix mask. The "S" matrix mask is generated by cyclically shifting an initial binary encoding sequence of a design. Each time, the sequence is shifted one position to the right, and multiple encoding patterns with cyclic structures are constructed in sequence. These patterns are then evenly distributed on different sectors of the rotating mask to form the "S" matrix mask, and the spacing between the patterns is equal.

6. The elderly dangerous actions monitoring system without video acquisition according to claim 5, characterized in that: The initial binary encoded sequence is generated by a pseudo-random method or a twin prime number algorithm.

7. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The housing (1) is also provided with an AC-DC regulated power supply module (8) and a control circuit module (9). The control circuit module (9) is provided with a drive circuit that controls the servo motor (6) to drive the encoding modulation disk (3) to rotate periodically. The AC-DC regulated power supply module (8) is used to supply power to the control circuit module (9) and the single pixel detector (4).

8. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The housing (1) has a threaded interface, and the industrial lens (2) is connected to the housing (1) through the threaded interface.

9. The elderly dangerous actions monitoring system without video acquisition according to claim 1, characterized in that: The servo motor (6) rotates at a speed in the centimeter range.

Images