A VAPE / smoke detection device and method
The 3D scanning and machine learning-based vape/smoke detection device addresses the unreliability of existing methods by accurately identifying vaping/smoking events in confined spaces with reduced false positives and negatives, ensuring privacy through anonymous detection.
Patent Information
- Application Number
- PCT/TR2024/050781
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-07-08
- Publication Date
- 2026-01-15
AI Technical Summary
Existing vape/smoke detection technologies in confined spaces suffer from high false positive and false negative rates, raising privacy concerns and health risks due to the reliance on cameras and chemical sensors, which are unreliable and invasive.
A 3D scanning-based vape/smoke detection device utilizing particulate matter sensors and processors to measure volume changes and particle concentrations, combined with depth mapping and machine learning algorithms to identify vaping/smoking events while ensuring privacy by maintaining anonymity.
Provides highly reliable detection of vaping/smoking activities with reduced false positives and negatives, protecting privacy by recognizing human presence and smoke/vapor sources without revealing identities.
Smart Images
Figure IMGF000009_0001 
Figure 00000016_0000 
Figure 00000017_0000
Abstract
Description
[0001] A VAPE / SMOKE DETECTION DEVICE AND METHOD
[0002] Technical Field of the Invention
[0003] The present invention relates to a device and a method for detecting vaping / smoking event in a confined space.
[0004] Background of the Invention
[0005] Like smoking, vaping also has become a big problem in confined spaces especially within schools and offices due to its harmful effects on people. Various techniques have been developed to detect or prevent vaping / smoking events. For instance, cameras or chemical detection sensors can be used to detect vaping / smoking. However, because of the privacy concerns, cameras cannot be used in private areas like restrooms, bathrooms, shower rooms, and hospital rooms. The reliability of vaping / smoking detection is essential for accurately identifying and preventing vaping / smoking in confined spaces. Current chemical detection sensors often result in significant high rates of false positives and false negatives when detecting vape aerosols in the air. This issue may lead to serious consequences for individuals falsely accused of vaping / smoking due to false positives. Thus, highly reliable vape / smoke detection that minimizes both false positives and false negatives is desirable in the state of the art.
[0006] Vaping / smoking is more common among young people and causes many health and environmental problems. The effects of vaping / smoking can also impact those in close proximity to the users. Therefore, detecting vaping / smoking activities in confined spaces is crucial for ensuring proper supervision and preventing harmful effects.
[0007] Brief Description of the Drawings
[0008] The accompanying drawings are given solely for the purpose of exemplifying an embodiment of the invention which will be explained in detail hereinafter:
[0009] Figure 1 is a schematic view of a vape / smoke detection device positioned inside a confined space during the first volume measurement in one exemplary embodiment of the present invention. Figure 2 is a schematic view of a vape / smoke detection device positioned inside a confined space wherein a human profile (potential vaper / smoker) is shown in one exemplary embodiment of the present invention. A depth map is depicted as gridlines. Figure 3 is a schematic view of a vape / smoke detection device positioned inside a confined space wherein a vaping / smoking profile depicting a human body with a cloud in the vicinity of head of the human body is shown in one exemplary embodiment of the present invention. A depth map is depicted as gridlines.
[0010] The elements illustrated in the figures are numbered as follows:
[0011] 100. Vape / smoke detection device
[0012] S. Confined space
[0013] V. Vaper / smoker
[0014] C. Vape / smoke cloud
[0015] Detailed Description
[0016] Embodiments of the present invention relates to a vape / smoke detection device (100) comprising at least one 3D scanning means for measuring a volume of a confined space; at least one particulate matter sensor for measuring floating particles in air; and a processor. The processor is adapted for: taking a first volume measurement of the confined space and a further volume measurement in a selected frequency of the confined space via the 3D scanning means, wherein if a volume measurement difference between the first volume measurement and any one of the further volume measurements exceeds a selected volume different threshold, measuring a distance between the particulate matter sensor and a position of occupied volume in the confined space causing the volume measurement difference via the 3D scanning means, taking particle measurement in the confined space, generating a vaping / smoking event signal if the particle measurement is inside a selected particle interval dependent on the distance (the distance between the particulate matter sensor and the occupied volume).
[0017] In the embodiment of the invention, the confined space should be understood as a closed or half-open space suitable for human entry including, but not limited to schools, classrooms, restrooms, storage areas, hospitals, hospital rooms, warehouses, cafeterias, offices, banks, elevators, financial institutions, government buildings and business entities.
[0018] In the embodiment of the invention, the particulate matter sensor might be a, but not limited to PM 2.5 sensor.
[0019] In the embodiment of the invention, the term "vape" herein states aerosol vapor from consuming products comprising but not limited heat-not-burn products, heated tobacco products, only heat tobacco products. The term "vaping" herein states consuming products, comprising but not limited heat-not-burn products, heated tobacco products, only heat tobacco products.
[0020] In this embodiment of the invention, the volume of the confined space is measured through the first volume measurement via the 3D scanning means. 3D scanning means is also called 3D range imaging means. The first volume measurement defines the volume confined space without any human (see figure 1). Then, the further volume measurements at a selected frequency are conducted to detect human entry over time. The preferred selected frequency is one volume measurement (for further volume measurements) every 10 to 200 seconds. Afterward, following each one of the further volume measurements, the processor checks the volume measurement differences between the first volume measurement (no human present in the confined space) and the further volume measurement (possible human entry to the confined space). If anyone of the volume measurement difference exceeds the selected volume different threshold, it is assumed that a human has entered the confined space. It is also assumed that confined space causing the volume measurement difference is realized by volume of human who has entered the confined space. The selected volume difference threshold may vary according to the age of the humans using the confined space. For example, If the confined space is an elementary school, the selected volume difference threshold may be slightly lower than the average volume of an elementary school student. If any one of the volume measurement difference / s exceeds a selected volume different threshold, distance between the particulate matter sensor and a position of occupied volume in the confined space causing the volume measurement difference is measured via the 3D scanning means. Thus, the distance between the particulate matter sensor and the human who has entered the confined space and is a potential vaper / smoker. Embodiments of the vape / smoke detection device (100) may comprise a casing covering the 3D scanning means and particulate matter sensor. The casing is suitable for mounting into the confined spaces. In this embodiment, distance between the particulate matter sensor and a position of occupied volume equals to range of the position of occupied volume to the vape / smoke detection device (100). Then, floating particles in the confined space is measured via particulate matter sensor. Exposure of the vaping / smoking to the particulate matter sensor mainly depends upon the distance between vaping / smoking source and the particulate matter sensor. In the presence of the vaping / smoking event, the particle measurement (the particle matter concentration) in the confined space should be inside a selected particle interval dependent on the distance between the particulate matter sensor and the position of occupied volume (human, the vaping / smoking source). Thus, if the particle measurement is inside a selected particle interval dependent on the distance between the particulate matter sensor and a position of occupied volume, a vaping / smoking event signal is generated via the processor. In this way, a highly reliable vape / smoke detection device (100) is realized.
[0021] Embodiments of the present invention also relates to a vape / smoke detection device (100) comprising at least one 3D scanning means for capturing a depth data of a confined space; at least one particulate matter sensor for measuring floating particles in air; and a processor. The processor is adapted for: generating a depth map of the confined space using a depth data captured by the 3D scanning means, identifying whether a human profile depicting a human body from the depth map is recognized, measuring a distance between the particulate matter sensor and a position of the human profile according to the confined space if the human profile is recognized, taking particle measurement in the confined space, generating a vaping / smoking event signal if the particle measurement is inside a selected particle interval dependent on the distance. In this embodiment of the invention, to detect human entry into confined space over time, the depth data of the confined space is captured via the 3D scanning means. By using the depth data, the depth map of the confined space is generated. Then, the processor checks whether the human profile is recognized from the depth map (see figure 2). The human profile is an object to be recognized identifying a human body (a potential vaper / smoker) who has entered the confined space. If the human profile is recognized, the distance between the particulate matter sensor and the position of the human profile according to the confined space. The human profile according to the confined space means that the position of the human in the confined space relative to the particulate matter sensor. Then, floating particles in the confined space is measured via particulate matter sensor. Exposure of the vaping / smoking to the particulate matter sensor mainly depends upon the distance between vaping / smoking source and the particulate matter sensor. In the presence of the vaping / smoking event, the particle measurement (the particle matter concentration) in the confined space should be inside a selected particle interval dependent on the distance between the particulate matter sensor and the position of human (the vaping / smoking source). Thus, if the particle measurement is inside a selected particle interval dependent on the distance between the particulate matter sensor and a position of human profile according to the confined space, a vaping / smoking event signal is generated via the processor. In this way, a highly reliable vape / smoke detection device (100) is realized. In this embodiment, if the depth map of the confined space is too dense, the number of points in the depth map may be reduced. A filter like voxel grid may applied the depth map to decrease the number of points in order to get more manageable data for depth map. Moreover, in order to facilitate human profile recognition, identify and extract key features from the depth map. The features might include geometric properties, surface normal, or shape descriptors like curvature and edges. Moreover, some descriptors may be calculated to provide a compact representation of the features. The descriptors might include spin images, point feature histograms (PFH), or fast point feature histograms (FPFH). The human profile recognition might be realized by using a pre-trained model or by using a database. The pre-trained model might be trained by using machine learning algorithms, deep learning with convolutional neural networks (CNNs) or 3D neural networks such as PointNet. The human profile recognition might be realized by matching the extracted features of the depth map against a database of known human profile features. For matching iterative closest point (ICP) or RANSAC might be used.
[0022] In one embodiment of the invention the selected particle interval is also dependent on the first volume measurement. Exposure of the vaping / smoking to the particulate matter sensor slightly depends upon the first volume measurement, in other words volume of the confined space. With this embodiment, a more precise and reliable detection of vaping / smoking might be achieved.
[0023] In one embodiment of the invention, the vape / smoke detection device (100) comprises a temperature sensor for measuring temperature of the confined space. In this embodiment, the selected particle interval is also dependent on the temperature of the confined space. Exposure of the vaping / smoking to the particulate matter sensor slightly depends upon temperature of the confined space. With this embodiment, a more precise and reliable detection of vaping / smoking might be achieved.
[0024] Embodiments of the present invention also relates to a vape / smoke detection device (100) comprising at least one 3D scanning means for capturing a depth data of a confined space; and a processor. The processor is adapted for: generating a depth map of the confined space using a depth data captured by the 3D scanning means, identifying whether a vaping / smoking profile depicting a human body with a cloud in the vicinity of head of the human body is recognized, generating a vaping / smoking event signal if the vaping / smoking profile is recognized from the depth map.
[0025] In this embodiment, to detect whether a vaper / smoker present in the confined, the depth data of the confined space is captured via the 3D scanning means. By using the depth data, the depth map of the confined space is generated. Then, the processor checks whether the vaping / smoking profile is recognized from the depth map (see figure 3). In this embodiment, the recognition of the vaping / smoking profile is a direct indicator for vaping / smoking event. The vaping / smoking profile is an object to be recognized identifying a human body with the vape / smoke cloud in the vicinity of head of the human body (a vaper / smoker). Thus, if the vaping / smoking profile is recognized from the depth map the vaping / smoking event signal is generated by the processor. In this way, a highly reliable vape / smoke detection device (100) is realized. In this embodiment, if the depth map of the confined space is too dense, the number of points in the depth map may be reduced. A filter like voxel grid may applied the depth map to decrease the number of points in order to get more manageable data for depth map. Moreover, in order to facilitate vaping / smoking profile recognition, identify and extract key features from the depth map. The features might include geometric properties, surface normal, or shape descriptors like curvature and edges. Moreover, some descriptors may be calculated to provide a compact representation of the features. The descriptors might include spin images, point feature histograms (PFH), or fast point feature histograms (FPFH). The vaping / smoking profile recognition might be realized by using a pre-trained model or by using a database. The pre-trained model might be trained by using machine learning algorithms, deep learning with convolutional neural networks (CNNs) or 3D neural networks such as PointNet. The vaping / smoking profile recognition might be realized by matching the extracted features of the depth map against a database of known vaping / smoking profile features. For matching iterative closest point (ICP) or RANSAC might be used.
[0026] One embodiment of the vape / smoke detection device (100) comprises at least one total volatile organic compound (tVOC) sensor for measuring volatile organic compounds in the confined space wherein the processor is adapted for generating a masking attempt event signal if the total volatile organic compounds sensor measurement exceeds a selected organic compound threshold. In some cases, vapers / smokers spray perfume, deodorant, and / or cologne in the confined space due to the fact that they think they can hide their vaping / smoking activities. The sensor detects the organic substances in these sprayed products in order to detect whether such a masking attempt is being made.
[0027] Table 1
[0028] Table 1 shows that the captured event in the confined space / generated event signal according to the particulate matter sensor (PM2p5(pg / m3)) value, tVOC sensor value, distance from source (m) (human, vaper / smoker, vape / smoke source) and volume of the confined space (m3) In this example, the selected particle interval for vaping / smoking source distance at 2 m distance and within 240 m3confined space volume is 100-500 pg / m3. During the captured / generated masking attempt event, the particulate matter sensor measurement (respectively 39 pg / m3, 30 pg / m3, 36 pg / m3) were not within the selected particle interval. Therefore, a vaping / smoking event signal was not been generated. However, since tVOC sensor measurement (respectively 113, 103, 140) exceeded selected organic compound threshold, a masking attempt event signal was generated. The selected organic compound threshold for vaping / smoking source distance at 2 m distance and within 240 m3confined space volume is 100.
[0029] One embodiment of the vape / smoke detection device (100) comprises at least one CO2sensor for measuring CO2value in the confined space. If the CO2sensor measurement exceeds a selected CO2level, it is verified that a human has entered the confined space.
[0030] In one embodiment of the vape / smoke detection device (100), the 3D scanning means comprises at least one time of flight sensor and a rotating means for rotating the time of flight (ToF) sensor.
[0031] In another embodiment of the vape / smoke detection device (100), the 3D scanning means comprises plurality of time of flight (ToF) sensor. ToF sensor uses pulses of infrared light to measure the distance between the scanner and the object such as the confined space for 3D scanning. The ToF sensor emits a short burst of light and records the time it takes for the light to bounce back from the object. By doing this for many points on the surface of the object, the scanner can create the depth map that represents the shape and size of the object in 3D space. The 3D scanning means comprising ToF sensor that can measure the phase difference between the emitted and reflected light. The phase difference is proportional to the distance between the sensor and the object, so by calculating the phase difference for each pixel on the sensor, the scanner can determine the depth of each point on the object. ToF scanners can capture depth data at high speed and resolution, making them suitable for scanning moving or dynamic objects In one embodiment of the vape / smoke detection device (100), the 3D scanning means comprises at least one light detection and ranging (LIDAR) sensor and a rotating means for rotating the light detection and ranging (LIDAR) sensor.
[0032] In another embodiment of the vape / smoke detection device (100), the 3D scanning means comprises plurality of light detection and ranging (LIDAR) sensor. Depth map might be a point cloud for the embodiment comprising the 3D scanning means having light detection and ranging (LIDAR) sensor.
[0033] In the embodiments of the invention, for recognition of the human profile or vaping / smoking profile, only a 3D depth map is generated, ensuring that the identity of the human (potential vaper / smoker or vaper / smoker) remains obscure, thereby protecting privacy. This enables the present invention to be used in confined spaces that require privacy such as toilets and restrooms.
[0034] Embodiments of the present invention also relates to a vape / smoke detection method comprising the steps of:
[0035] - taking a first volume measurement of a confined space and a further volume measurement in a selected frequency of the confined space via at least one 3D scanning means;
[0036] - if a volume measurement difference between the first volume measurement and any one of the further volume measurements exceeds a selected volume different threshold, measuring a distance between the particulate matter sensor and a position of occupied volume in the confined space causing the volume measurement difference via the 3D scanning means;
[0037] - taking particle measurement in the confined space via at least one particulate matter sensor for measuring floating particles in air;
[0038] - if the particle measurement is inside a selected particle interval dependent on the distance, generating a vaping / smoking event signal.
[0039] Embodiments of the present invention also relates to a vape / smoke detection method comprising the steps of:
[0040] - capturing a depth data of a confined space via at least one 3D scanning means; - generating a depth map of a confined space using the depth data of the confined space;
[0041] - identifying whether a vaping / smoking profile depicting a human body from the depth map is recognized;
[0042] - if the vaping / smoking profile is recognized, measuring a distance between the particulate matter sensor and a position of the vaping / smoking profile in the confined space;
[0043] - taking particle measurement in the confined space via at least one particulate matter sensor for measuring floating particles in air;
[0044] - if the particle measurement is inside a selected particle interval dependent on the distance, generating a vaping / smoking event signal.
[0045] In one embodiment of the vape / smoke detection method, the selected particle interval is also dependent on the first volume measurement.
[0046] An embodiment of the vape / smoke detection method, comprises a step of measuring temperature of the confined space via a temperature sensor. In this embodiment, the selected particle interval is also dependent on the temperature of the confined space.
[0047] Embodiments of the present invention also relates to a vape detection / smoke method comprising the steps of:
[0048] - capturing a depth data of a confined space via at least one 3D scanning means;
[0049] - generating a depth map of a confined space using the depth data of the confined space;
[0050] - identifying whether a vaping / smoking profile depicting a human body with a cloud in the vicinity of head of the human body is recognized;
[0051] - if the vaping / smoking profile is recognized from the depth map, generating a vaping / smoking event signal.
Claims
CLAIMS1. A vape / smoke detection device (100) comprising at least one 3D scanning means for measuring a volume of a confined space; at least one particulate matter sensor for measuring floating particles in air; and a processor adapted for: taking a first volume measurement of the confined space and a further volume measurement in a selected frequency of the confined space via the 3D scanning means, wherein if a volume measurement difference between the first volume measurement and any one of the further volume measurements exceeds a selected volume different threshold, measuring a distance between the particulate matter sensor and a position of occupied volume in the confined space causing the volume measurement difference via the 3D scanning means, taking particle measurement in the confined space, generating a vaping / smoking event signal if the particle measurement is inside a selected particle interval dependent on the distance.
2. A vape / smoke detection device (100) comprising at least one 3D scanning means for capturing a depth data of a confined space; at least one particulate matter sensor for measuring floating particles in air; and a processor adapted for: generating a depth map of the confined space using a depth data captured by the 3D scanning means, identifying whether a human profile depicting a human body from the depth map is recognized, measuring a distance between the particulate matter sensor and a position of the human profile according to the confined space if the human profile is recognized, taking particle measurement in the confined space, generating a vaping / smoking event signal if the particle measurement is inside a selected particle interval dependent on the distance.
3. A vape / smoke detection device (100) according to any of the preceding claims wherein the selected particle interval is also dependent on the first volume measurement.
4. A vape / smoke detection device (100) according to any of the preceding claims comprises a temperature sensor for measuring temperature of the confined space wherein the selected particle interval is also dependent on the temperature of the confined space.
5. A vape / smoke detection device (100) comprising at least one 3D scanning means for capturing a depth data of a confined space; and a processor adapted for: generating a depth map of the confined space using a depth data captured by the 3D scanning means, identifying whether a vaping / smoking profile depicting a human body with a cloud in the vicinity of head of the human body is recognized, generating a vaping / smoking event signal if the vaping / smoking profile is recognized from the depth map.
6. A vape / smoke detection device (100) according to any of the preceding claims comprising at least one total volatile organic compounds sensor for measuring volatile organic compounds in the confined space wherein the processor is adapted for generating a masking attempt event signal if the total volatile organic compounds sensor measurement exceeds a selected organic compound threshold.
7. A vape / smoke detection device (100) according to any of the preceding claims wherein the 3D scanning means comprises at least one time of flight sensor and a rotating means for rotating the time of flight sensor.
8. A vape / smoke detection device (100) according to any of the preceding claims wherein the 3D scanning means comprises plurality of time of flight sensor.
9. A vape / smoke detection device (100) according to any of the preceding claims wherein the 3D scanning means comprises at least one light detection and ranging sensor and a rotating means for rotating the light detection and ranging sensor.
10. A vape / smoke detection device (100) according to any of the preceding claims wherein the 3D scanning means comprises plurality of light detection and ranging sensor.
11. A vape / smoke detection method comprising the steps of:- taking a first volume measurement of a confined space and a further volume measurement in a selected frequency of the confined space via at least one 3D scanning means;- if a volume measurement difference between the first volume measurement and any one of the further volume measurements exceeds a selected volume different threshold, measuring a distance between the particulate matter sensor and a position of occupied volume in the confined space causing the volume measurement difference via the 3D scanning means;- taking particle measurement in the confined space via at least one particulate matter sensor for measuring floating particles in air;- if the particle measurement is inside a selected particle interval dependent on the distance, generating a vaping / smoking event signal.
12. A vape / smoke detection method comprising the steps of:- capturing a depth data of a confined space via at least one 3D scanning means;- generating a depth map of a confined space using the depth data of the confined space;- identifying whether a vaping / smoking profile depicting a human body from the depth map is recognized;- if the vaping / smoking profile is recognized, measuring a distance between the particulate matter sensor and a position of the vaping / smoking profile in the confined space;- taking particle measurement in the confined space via at least one particulate matter sensor for measuring floating particles in air;- if the particle measurement is inside a selected particle interval dependent on the distance, generating a vaping / smoking event signal.
13. A vape / smoke detection method according to claim 10 or claim 11, wherein the selected particle interval is also dependent on the first volume measurement.
14. A vape / smoke detection method according to any of claim 10-12, comprising the steps of - measuring temperature of the confined space via a temperature sensor wherein the selected particle interval is also dependent on the temperature of the confined space.
15. A vape / smoke detection method comprising the steps of:- capturing a depth data of a confined space via at least one 3D scanning means;- generating a depth map of a confined space using the depth data of the confined space;- identifying whether a vaping / smoking profile depicting a human body with a cloud in the vicinity of head of the human body is recognized;- if the vaping / smoking profile is recognized from the depth map, generating a vaping / smoking event signal.