MINIATURISED STEREOSCOPIC THERMAL SENSOR FOR AN AUTOMATIC COUNTING DEVICE

Abstract

The subject matter of the present invention concerns a sensor for counting and/or determining the direction of passage of objects and/or living beings, each having a thermal signature, comprising or consisting of: a pyroelectric component, preferably digital, integrating at least two pairs of detection cells; and one cell of which is masked in each pair of cells.

The present invention further concerns a mechanical passage collecting unit comprising such a sensor and an optical lens. The present invention further concerns the use of such a sensor, for example in the form of a mechanical passage collecting unit, in a device. The present invention concerns the device as such. The present invention also concerns the sensor data processing algorithm.

Claims

1. A sensor for counting and/or determining the direction of passage of objects and/or living beings, each having a heat signature, comprising: a pyroelectric component, preferably digital, integrating at least two pairs of detection cells; and one cell of which is masked in each pair of cells.

2. The sensor according to claim 1, wherein each pair of cells comprises a masked reference cell and an unmasked detection cell.

3. The sensor according to claim 2, wherein the reference cells are side by side.

4. The sensor according to claim 2, wherein the masking of the reference cells consists of a single continuous piece.

5. A mechanical passage collecting unit, comprising: at least one sensor according to claim 1; and at least one optical lens arranged in front of the pyroelectric component of said sensor.

6. The mechanical passage collecting unit according to claim 5, wherein the optical lens is adapted to the thermal infrared signal.

7. The mechanical passage collecting unit according to claim 5, wherein the optical lens is a Fresnel lens, a Germanium lens or a zinc selenide lens.

8. The mechanical passage collecting unit according to claim 5, wherein the detection cells of the said sensor are in the focal plane of the optical lens or the lens and the sensor are adjusted in such a way that the cells of the sensor are close to the focal plane of the lens.

9. A device for counting and determining the direction of passage of objects and/or living beings each having a heat signature, wherein said device comprises at least one sensor according to claim 1.

10. The device as claimed in claim 9, wherein said at least one sensor according to claim 1 is integrated into a mechanical passage collecting unit having at least one said sensor and at least one optical lens arranged in front of the pyroelectric component of said sensor.

11. The device according to claim 9, wherein the operation of the sensor is stereoscopic.

12. The device according to claim 9 further comprising: a power supply means; optionally a marking specifying the upper side of the device and optionally a marking indicating the detection field of said device and/or the direction of passage of said objects and/or living beings; and at least one processing unit adapted to determine the number and/or to determine the direction of passage of each object and/or living beings having a thermal signature passing in front of said device.

13. The device according to claim 12, wherein said at least one sensor is integrated into a passage collecting unit having at least one said sensor and at least one optical lens arranged in front of the pyroelectric component of said sensor.

14. The device according to claim 12, wherein said device comprises at least one data transmission means.

15. The device according to claim 12, wherein said device comprises at least one data storage means.

16. The device according to claim 12, wherein said device can also comprise at least one magnetometer.

17. An algorithm for implementing a sensor according to claim 1, wherein the following steps are applied by said algorithm: a first step (a) where all the detection variables are initialized to zero, including the number of passages; where applicable, a step (b) corresponding to the acquisition and processing of the data generated during the passage of an object and/or a living being having a thermal signature in front of the sensor with a positive assignment or negative of a passage unit by passage of an object and/or a living being having a thermal signature in front of said sensor; optionally, a step of comparison of the absolute data and the derivatives of two curves of similar shapes and coming from the same sensor; a step (c) of adding or subtracting from the number of passages in memory of a unit according to the assignment of step (b); where appropriate, a step (d) of reproduction of steps (b) and (c); and optionally at least one step (e) of data storage with optionally a resetting to zero of the variables at each storage.

18. The algorithm according to claim 17, wherein the said algorithm is integrated into a mechanical passage collecting unit having at least one said sensor and at least one optical lens arranged in front of the pyroelectric component of said sensor, or of a device for counting and determining the direction of passage of objects and/or living beings, each having a heat signature having at least one said sensor.

Description

FIGURES

[0063] FIG. 1 represents a device (D) according to the present invention comprising a sensor (C) according to the present invention, a lens (L), a processing unit (TT) and a transmission unit (TS). The direction of passage of the moving thermal targets is represented by the arrows (F), and the device is arranged so that this passage crosses the detection field (CH). The thermal targets emit infrared signals (SIR) which are detected by the sensor (C) after their passage through the lens (L). The sensor then sends an analog or digital signal (S1) to the processing unit (TT) which processes the signal so that it can be used by the transmission unit (TS)—and/or storage if applicable. The transmission unit (TS) transmits the useful information according to its designated protocol. In a preferred manner, the device comprises a single sensor (IR) according to the present invention (comprising for example 2×2 detection cells, such as 2×2 pixels), placed in the focal plane or close to it of the lens (L), which can be an IR lens (Fresnel, germanium or even zinc selenide lens, for example).

[0064] FIG. 2 represents a sensor (C) according to the present invention, comprising 2×2 detection cells A/A′ and B/B′. Four embodiments are represented below in FIG. 2A, FIG. 2B, FIG. 2C and FIG. 2D. In FIGS. 2A and 2C, the pairs of detection cells are arranged so that the reference cells (A′ and B for 2A, or A and B′ for 2C) are not one next to each other, whereas in FIGS. 2B and 2D, the reference cells (A′ and B for 2B, or A′ and B′ for 2D) are next to each other. In the latter cases, the hashed-greyed out masking element (M) is in one piece (allowing easy positioning and optimized masking). In FIG. 2B, the pairs A/A′ and B/B′ are thus said to be “crossed”. In FIGS. 2A and 2C, two masking elements (M) are used. Thus, in all the cases of figures represented, the sensor can be pyroelectric integrating two pairs of detection cells, of which two masked cells serve as reference, and the two detection cells (not masked) allow a stereoscopic type of algorithm. It is important that for each detection cell, the remaining (masked) cell serves as a reference in order to have a pair-to-pair comparison basis.

[0065] FIG. 3 represents a graph of signal intensity (in ordinates—arbitrary scale of values) in function of time (in milliseconds) when a pedestrian (slow passage) passes in front of the sensor according to the present invention (sensor according to Example 1 below). The blue and red curves respectively represent the signals obtained for a first pair of cells and a second pair of detection cells used in the sensor according to the present invention. The green curve represents the events (in the form of peaks), for example when one of the values of the blue and/or red curves reaches a threshold value (determined by the user or set by default), or when the processor has finalized an event.

[0066] FIG. 4 represents a graph of signal strength (in ordinate—arbitrary scale of values) versus time (in milliseconds) when a vehicle such as a bicycle (fast passing) passes in front of the sensor according to the present invention (sensor according to Example 1 below). The legend and comments for FIG. 3 are applicable to FIG. 4.

EXAMPLES

Example 1: Created Device

[0067] A device according to the present invention was created by using the following parts: [0068] Box: ABS casing of dimensions 101×77×40; [0069] Weak central unit: card is built around the “MSP430F248®” microcontroller; [0070] PYQ5848 digital matrix thermal sensor with two pairs of crossed AA′ and BB′ detectors. The sensor is mechanically implanted independently of the electronic board; the connection with the card is made by a wired connection of three wires; [0071] Masking by concealment of two thermal pixels B & A′ thanks to a single rectangular sticker. The two hidden pixels become the thermal references of the two remaining detectors. The two detectors provide similar thermal curves which are then compared over time using their derivatives; [0072] Battery: CR123 NiMnO2, voltage of 3V, advertised capacity of 1500 mAHrs; [0073] Sigfox transmission card: “SiPy®” cards (from the PyCom® manufacturer); [0074] Internal storage of quarter-hour counting results: “I2C 24C1025®” memory; [0075] Switch: electronic card equipped at two ends with Hall effect electromagnetic sensors.

[0076] The various elements listed above are simply assembled to obtain a device according to the present invention.

Example 2: Tested Device

[0077] Devices according to example 1 were tested according to the following protocols: [0078] Scenario passage tests at one meter, four meters and seven meters with single passages and double passages in both directions. Video sequences associated with algorithmic data measurements were recorded using KDLINK® software for replay and analysis. Usage adjustments of the algorithm resulted in an accuracy of +/−5% on scenarios of single and grouped passages (excluding stagnation and masking). [0079] Field tests were carried out repeatedly at the entrance/exit of the “Village Gaulois de Pleumeur-Bodou” (tourist park welcoming approximately 60,000 visitors per year), where a device according to example 1 was concealed and positioned at the entrance to the park. The obtained figures by the devices according to example 1 were compared with reference devices and with the number of paid admissions to the park.

[0080] Thus, on Aug. 15, 2018, for 1192 paying entries, the device according to example 1 recorded 1248 outward and return passages. The next day, for 921 paying entries, the sensor counted 990 outward and return passages. These figures are consistent with the fact that the sensor counts all passages and aims for an accuracy of +/−10%.

Example 3: Determination of the Nature and Direction of Passage of the Object

[0081] A first method was applied by adjusting the algorithm of the counting boxes to calculate and estimate the speed of passage of the detected targets and to classify them according to the speed the nature of the object.

[0082] Thus, the stereoscopic effect of the sensors gives types of curves of the same order.

[0083] The order of the curves makes it possible to determine the direction of passage. For example, in FIGS. 3 and 4, when the red curve (and cross) is followed by a blue curve (and circles), i.e., the peak of the blue curve appears after the peak of the red curve according to the time scale of the abscissa (on the right side of the graph here), this corresponds to a determined direction (an entry for example). When the red/blue order is reversed (i.e., the peak of the red curve is to the right of the peak of the blue curve), then this corresponds to the opposite direction. Each of these curves (red or blue) corresponds to the signal received by each pair of cells (one cell masked and the other not) sequentially.

[0084] The detection of the extrema is done at the passage to 0 of the two curves (see peaks on the green curves (and triangles) of the events). It is then a question of measuring the difference between the extrema to find out a speed of passage allowing to classify the type of practice. A small-time variation typically corresponds to the signal left by a vehicle such as a bicycle. The threshold can be adjustable for a variation of time left by a pedestrian. The graph in FIG. 3 is representative of the passage of a pedestrian and the graph of FIG. 4 is representative of the passage of a vehicle such as a bicycle.

[0085] Concretely, scenario tests were carried out to develop and then validate an estimate of the speed of passage. A software allowing to visualize the curves of the two sensors for each acquisition sequence and to associate to it control video sequences was used. Here, the software allowing this in the tests carried out is called “KDLINK”. During these scripted tests, it was possible to reliably and repetitively classify pedestrians and bicycles, while indicating their direction of passage.

[0086] The adjustment of the sensor algorithm has been validated in the field in many cases, including a counting in Houat (France) in September 2019. This counting involved ten boxes for a full week.

[0087] A second method for determining the nature of the object passing in front of the sensor is to couple said sensor with a magnetometer making it possible to detect the passage of a metallic mass close to the sensor and to implement the management of the measurement of the magnetic sensor in the algorithm in order to specify that it is an object generating or sensitive to a magnetic field. This alone can make it possible to distinguish passages of cars in order to count only bicycles or pedestrians. This method can be combined with the first method above (from example 3), and depending on the context, one or the other method can then be favored when the data is contradictory.

[0088] In order not to compromise too much the autonomy of the device/sensor according to the present invention by the addition of a magnetometer, a particular embodiment was carried out in which the measurement period of the magnetometer was adapted in the following way: [0089] by default, the measurement period is 500 milliseconds (consumption divided by 25) and gives the signal reference; [0090] during thermal signal detections, this period temporarily passes to 20 milliseconds and allows metal detection to be well characterized until the end of passage detection.

[0091] A generalization can thus be made as follows: [0092] by default, the measurement period is from 50 milliseconds to 2 seconds, for example from 100 milliseconds to 1000 milliseconds, preferably from 200 milliseconds to 750 milliseconds such as 500 milliseconds, and gives the signal reference; [0093] during thermal signal detections, this period momentarily passes within a range of 10 to 100 milliseconds, preferably 20 or 40 milliseconds, which makes it possible to properly characterize the metal detection until the end of passage detection.

[0094] In conclusion, thanks to the paired management of the two sensors (the one according to the present invention and the magnetometer), the boxes according to the present invention remain energy autonomous and can classify the passages of vehicles (for example bicycle/car/motorcycle/truck) vs. people (e.g., bicycle/pedestrian/scooter/rollerblade on shared country roads or on cycle lanes (for bicycles)).