JP2003109128A - Intruder detecting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an intruder detecting device which exactly detects an intruder. SOLUTION: The intruder detecting device is provided with optical sensor arrays 12a and 12b outputting image signals from the arrays in positions corresponding to the formed image of a pair of image forming lenses 11a and 11b, a signal processing part 13 converting the image singles into image data, a distance detecting circuit 14 calculating distance data from image data to the intruder, a background distance data generating means 21 generating background distance data, a differential distance data generating means 22 generating differential distance data, a calculating means 23 calculating the detected height of an object based on differential distance data and an intruder judging means 24 judging whether an object is an intruder or not based on detected height.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an intruder detecting device for detecting an intruder in a facility or a room.

[0002]

2. Description of the Related Art Conventionally, an intruder detection device for detecting an intruder that has invaded a facility or a room has been developed, but various principles of detecting an intruder have been adopted. For example,
As conventional techniques, (A) a detection method using a pyroelectric sensor that detects infrared rays emitted from a person, (B) a detection method that uses an opposed infrared sensor that detects light emitted from a light emitting means by a light receiving means, etc. Can be mentioned.

Further, (C) a light beam is projected toward an intruder, and reflected light from the intruder is PSD (Position Sen).
The invention of the distance-measuring intruder detector, which detects the distance information to the intruder by the principle of triangulation and detects the intruder by using this distance information, is published. (JP-A-5-81575
No. 5,172,515).

Further, as an example in which a detection principle different from the above (A) to (C) is disclosed, which does not detect an intruder, (D) obtain an image of an occupant seated in a passenger seat. , Multiple parts of the occupant (chest
An invention has been disclosed (Japanese Patent Application Laid-Open No. 11-217056) regarding an attitude determination device that obtains distance information of the waist and thighs and determines the attitude of an occupant using this distance information.

Regarding the principle of distance measurement by the distance measurement processing device using triangulation in this posture determination device,
The principle will be described with reference to the distance measurement principle explanatory diagram of FIG. As shown in FIG. 14, the horizontal axis X and the vertical axis Y are set with the center of the imaging lenses 201 and 202 as the origin O, and the imaging positions by the respective lenses are L ₁ and R ₁ . The coordinates of the center point O _L of the imaging lens 201 are (−B / 2, 0).
In, the coordinates of the center point O _R of the imaging lens 202 (B / 2,
0). The coordinates of the point M of the object (subject) are (-x,
y), a position where a straight line passing through the point O _L and parallel to the straight line MO intersects with the photosensor array 211 is L ₀ , and a position where a straight line passing through the point O _R and parallel to the straight line MO intersects with the photosensor array 212 is _denoted as R ₀ . To do.

Here, a _L represents the distance between the points L ₀ and L _1, and a _R represents the distance between the points R ₀ and R ₁ . L ₀ and R ₀ are reference positions for obtaining a _L and a _R. At this time, △ MO _L O and _{△ O L L 1 L 0,} △ MO R
Since O and ΔO _R R ₁ R ₀ are similar to each other, they can be expressed by the following equation.

[0007]

[Equation 1] (-x + B / 2) f = a _L · y

[0008]

(2) (x + B / 2) f = a _R · y

The following equation can be obtained by arranging these equations 1 and 2 so as to eliminate x.

[0010]

[Formula 3] y = B · f / (a _L + a _R )

When the sensor pitch is p and the number of sensors is N, the following equation is obtained.

[0012]

[Formula 4] a _L + a _R = p · N

The equation (4) can be transformed into the following equation.

[0014]

[Formula 5] y = B · f / (p · N)

[0015] That is, the distance a _L between the imaging position L ₁ and the point L ₀ of the left photosensor array 211, and the distance between the imaging position R ₁ and the point R ₀ of the right photosensor array 212 a _R Therefore, the distance y to the object can be obtained from the above equation 3. The distance measuring principle itself is well known.

In order to determine the relative positions of the two images as described above, the distance calculation processing device performs, for example, correlation calculation as described below. The output of each element of both photosensor arrays 211 and 212 is converted into, for example, an 8-bit digital signal and then stored in a register as a memory (not shown). From the stored data,
A pair of calculation regions 221, 222 as shown in the correlation calculation explanatory diagram of FIG. 15 is selected by a partial data extraction unit (not shown). This calculation regions 221 and 222 is composed of n elements as shown in FIG. 15, respectively _A 1 to A
_n , B _{1 to} B _n quantized data.

Here, a predetermined angle from the front of the sensor (see FIG.
4, the distance index (a _L + a _R ) to the object at the angle between the Y axis, the object coordinate M, and the straight line MO connecting the origin O)
To obtain the above, the correlation detection / distance calculator not shown for the quantized data sets windows W _{1 to} W _{m + 1} of a predetermined size as shown in FIG. 16 and shifts them alternately by one array unit (1 bit). (M + 1) sets of subsets C _{0 to} C _m are considered, and evaluation functions f (C ₀ ) to f (C ₀ are formed of sums of absolute values of differences in quantized data for each of the subsets.
_m ) and obtain the combination Ck that minimizes this evaluation function value, the degree of displacement of the left and right images is known from the value of the subscript k, and as shown in the above mathematical expression 3 (a _L + a _R ).
It is known that a distance index proportional to is required (for example, Japanese Patent No. 2676985, corresponding US Patent No. 5).
602944 or corresponding German Patent No. 4121
145). Therefore, for example, the calculation areas 221 and 222 as shown in FIG.
While sequentially shifting by bits (1 pitch: p),
By performing the process of obtaining the distance index as described above for each of the calculation areas 221, 222, s discrete distance data can be obtained from the first to sth calculation areas. As the prior art, the ones mentioned in the above (A) to (D) are known.

[0018]

However, the method of detecting an intruder as described above has various problems.
The method using the pyroelectric sensor of (A) has a problem that many false detections occur because small animals (for example, a cat, a mouse, etc.) that emit infrared rays are detected in addition to humans. In the detection method using the opposed infrared sensor of (B), since the blocking of the infrared rays is detected, there is a problem that the intrusion cannot be detected when an intruder passes through or jumps over the infrared rays.

(C) JP-A-5-81575 and JP-A-5-815
In the range-finding type intruder detector disclosed in No. 172515, since a light beam is projected onto the intruder and the reflected light from the intruder is detected, when the distance to the intruder is long, or
When the reflectance of the intruder is low, there is a problem that the amount of the reflected light reflected from the intruder is insufficient and the distance to the intruder cannot be obtained.

Further, (D) the occupant's posture determination device disclosed in Japanese Patent Laid-Open No. 11-217056 has a very excellent detection performance by using the distance measurement principle of triangulation, but is intruded. In order to measure a person, there is a problem in that it cannot be simply applied because it is expected that a small animal other than a human will be detected by this distance measuring principle. There is a need for an intruder detection device that uses a new ranging principle that can reliably detect an intruder.

Further, as a problem common to these (A) to (D), there is a problem that the influence of background objects other than intruders such as furniture and pillars in the image cannot be avoided.
There has been a demand for avoiding the influence of these background objects and improving the detection accuracy.

The present invention has been made to solve the above problems, and an object of the present invention is to provide an intruder detection device for accurately detecting an intruder.

[0023]

In order to solve the above-mentioned problems, the invention of claim 1 is such that a pair of image-forming devices are arranged so that their optical axes are parallel to each other and are imaged on the same image-forming plane. A lens, and an optical sensor array that is arranged on an image forming plane of the pair of image forming lenses and outputs an image signal from an array at a position corresponding to image formation by each of the pair of image forming lenses,
Based on the distance data output from the distance detection means, distance detection means for measuring the distance to the object in the image based on the image signal output from the optical sensor array to create and output distance data A background distance data creating unit that creates and outputs background distance data that is the distance to the background of the object, and obtains the difference between the background distance data and the latest distance data, and when the difference value exceeds a predetermined value. The difference distance data creating means for outputting the latest distance data as difference distance data, the difference distance data output from the difference distance data creating means, and the intersection angle between the visual field direction of the optical sensor array and the vertical direction with respect to the ground. Viewing direction angle,
And a calculating means for calculating the detected height of the object using the mounting height of the optical sensor array respectively, and whether or not the object is an intruder based on the detected height calculated by the calculating means. Intruder determining means for determining is provided.

According to a second aspect of the present invention, in the intruder detecting apparatus according to the first aspect, the intruder determining means is configured to detect the detected height of the object higher than a predetermined determination height. Determines that the object is an intruder.

According to the invention of claim 3, in the intruder detection device according to claim 1 or 2, the optical sensor arrays are formed with different heights so as to have different viewing direction angles. A plurality of pairs of photosensor arrays arranged on a plane, wherein the plurality of pairs of photosensor arrays detect image signals of an object at different viewing direction angles, and the distance detection means, the background distance data creation means, It is characterized in that the difference distance data creating means, the calculating means, and the intruder determining means determine intruders in different visual field directions.

According to a fourth aspect of the present invention, in the intruder detection device according to any one of the first to third aspects, the optical sensor array is an array divided into a plurality of columns, It is characterized in that an image signal output from one measurement window defined by a plurality of some arrays among all the arrays is used.

Further, the invention of claim 5 is the intruder detecting apparatus according to any one of claims 1 to 4, wherein the background distance data creating means is an average value of distance data measured a plurality of times. Is used as background distance data.

[0028] According to a sixth aspect of the invention, in the intruder detecting apparatus according to the fifth aspect, the background distance data creating means uses the distance data obtained a plurality of times after the intruder is not detected. Is characterized by.

Further, the invention of claim 7 is the intruder detection device according to any one of claims 1 to 6, wherein the difference distance data creating means is provided for each measurement window on each sensor line. For each distance data, when the difference between the background distance data and the latest distance data exceeds the judgment distance, the latest distance data is substituted into the difference distance data, and the difference between the background distance data and the latest distance data is the judgment distance. It is a means for substituting the invalid distance data into the difference distance data in the following cases.

Further, the invention of claim 8 is the intruder detecting device according to claim 7, wherein the intruder determining means omits the determination when the previous difference distance data is invalid distance data, It is characterized in that it is a means for judging in the case of the latest distance data.

Further, the invention of claim 9 is the background for updating the background distance data output from the background distance data creating means in the intruder detecting device according to any one of claims 1 to 8. It is characterized by comprising distance data updating means.

[0032]

BEST MODE FOR CARRYING OUT THE INVENTION A first embodiment of an intruder detection device of the present invention will be described below with reference to the drawings. In the drawings, the same reference numerals indicate the same or corresponding parts. FIG. 1 is a schematic configuration diagram of the first embodiment of the present invention. The intruder detection device includes a distance sensor 10 and a microcomputer 2.
It has at least 0. Further, the distance sensor 10
Includes an imaging lens 11, an optical sensor array 12, a signal processing unit 13, and a distance detection circuit 14. In the present embodiment, the distance detecting means of the present invention corresponds to the signal processing unit 13 and the distance detecting circuit 14.

Further, the microcomputer 20 comprises a background distance data creating means 21, a difference distance data creating means 22, a calculating means 23, and an intruder judging means 24 as a program. Distance data is transmitted from the distance sensor 10 to the microcomputer 20.

FIG. 2 is a detailed configuration diagram of the distance sensor of this embodiment, and FIG. 3 is a layout diagram of the optical sensor array. The imaging lens 11 is a pair of imaging lenses 11 as shown in FIG.
a and 11b, the optical axes are parallel to each other, and the base length of the optical axis is B. The optical sensor array 12 is
For example, it is a CCD linear array sensor, and in the present embodiment, as shown in FIG. 3, a pair of photosensor arrays 12a and 12b are provided in the upper line and the photosensor array 1 is provided in the middle line.
2c and 12d have optical sensor arrays 12e and
12f are arranged.

The optical sensor arrays 12a to 12f on the upper, middle, and lower lines are formed by the imaging lenses 11a and 11 respectively.
It is arranged so as to have a focal length f with respect to b, and converts the image of the object 100 formed by the imaging lenses 11a and 11b into image signals a and b, and outputs the image signals to the signal processing unit 13. To do. Since the processing from the signal processing unit 13 is the same, the optical sensor arrays 12c,
The description of the optical sensor arrays 12e and 12f on the lower line 2d and 2d will be omitted, and the optical sensor arrays 12a and 12b on the upper line will be described below as representatives.

The signal processor 13 includes amplifiers 15a and 15a.
b, A / D converters 16a and 16b, and a storage device 17. Image signals a and b from the photosensor arrays 12a and 12b are amplified by amplifiers 15a and 15b, converted into digital data by A / D converters 16a and 16b, and output to the storage device 17 as image data a and b. It

The distance detection circuit 14 is a circuit composed of a microcomputer, and calculates the distance to the object 100 by the triangulation principle using the left and right image data a and b stored in the storage device 17. It is output to the outside as distance data. Although not shown in the figure, the remaining two pairs of optical sensor arrays 12c and 12d and optical sensor array 12 are shown.
Similarly, e and 12f are also connected to the signal processing unit 13, and the distance data obtained in the same manner are output to the external microcomputer 20. Based on the distance data, the background distance data creating means 21, the differential distance data creating means 22, the calculating means 23, and the intruder determining means 24 in the microcomputer 20 determine whether or not the object is an intruder according to the principle described later. Is determined.

Next, an installation mode of the intruder detection device will be described. FIG. 4 is a perspective view illustrating the installation of the intruder detection device installed indoors, and FIG. 5 is a side view of the same. 4 and 5, the distance sensor 10 is attached to the wall 1a of the room 1 or the like and arranged so as to detect an object (intruder) 100. As is apparent from FIG. 5, which is a side view seen from the horizontal direction A side shown in FIG.
Is arranged at a mounting height H _s so that the optical axis direction thereof is tilted at an installation angle θ _s with respect to a horizontal direction B that is parallel to the reference ground surface and looks down. .

The three pairs of photosensor arrays described above each have a measurement field of view surrounded by the respective regions shown in the measurement fields of view 3 to 5 in FIGS. 4 and 5. That is, the measurement field of view 3 of the photosensor arrays 12a and 12b in the upper line of the photosensor array 12 is a region determined by the alternate long and short dash line.
Of the photosensor arrays 12c and 12d in the middle line
The measurement field of view 4 is a region surrounded by a solid line, and the measurement field 5 of the photosensor arrays 12e and 12f in the lower line of the photosensor array 12 is a region surrounded by a dotted line. All of these are measurement fields of view that look down on the object from above.

Next, the principle of detecting an intruder will be described.
First, a method of calculating distance data by the distance sensor will be described. Although there are 3 pairs of optical sensor arrays,
Since the pair operates on the same principle, the optical sensor arrays 12a and 12b in the upper line are shown here.
A method of calculating the distance data will be described by taking the pair of as an example.

FIG. 6 is an explanatory view of the principle of distance detection. Each conclusion
The origin O is the midpoint of the base length B of the image lenses 11a and 11b.
The horizontal axis X and the vertical axis Y, and the point where the object 100 is located
The coordinates of M are (-x, y). In this case
Image forming position L₁, R₁The coordinates of (-B / 2-
a_L1, -F), (B / 2 + a_R1, -F), imaging lens
Center point O of Z 11a_LCoordinates are (-B / 2, 0), image formation
Center point O of lens 11b_RCoordinates are (B / 2, 0), point
The coordinates of the intersection point N between the perpendicular line drawn from M to the X axis and the X axis are
(-X, 0), point O_LDown to the optical sensor array 12a
Position L of vertical line ₀Coordinates are (-B / 2, -f), point O
_RPosition R of the vertical line from the optical sensor array 12b₀
The coordinates are (B / 2, -f).

At this time, ΔMO _L N and ΔO _L L ₁ L ₀ , Δ
Since MO _R N and △ _{O R} R 1 R ₀ is similar respectively,
The following formula is established.

[0043]

(6) (−x + B / 2) f = (a _L1 ) y

[0044]

## EQU7 ## (x + B / 2) f = (a _R1 ) y

The following equation can be obtained by rearranging x from equations 6 and 7 so as to eliminate them.

[0046]

Y = B · f / (a _L1 + a _R1 )

As described above, if the x-coordinates a _L1 and a _R1 of the image forming positions L ₁ and R ₁ are known, the distance L to the object 100 is L.
Can be calculated. (Principle of triangulation)

In the signal processing unit 13, the image forming positions L ₁ and R ₁
The image data indicating the x-coordinates a _L1 and a _R1 of is measured. Specifically, as shown in FIG. 2, the optical sensor array 12a,
Image signals a and b from 12b are amplified by amplifiers 15a and 15b, converted into digital data by A / D converters 16a and 16b, and then stored in the storage device 1 as image data a and b.
7 is output. The image data a and b stored in the storage device 17 are output to the distance detection circuit 14.

Next, distance measurement using the image data a and b will be described. FIG. 7 is an operation explanatory diagram of the distance detection circuit 14. The distance detection circuit 14 has a structure as shown by the solid line in FIG.
Image data a, b on the photo sensor arrays 12a, 12b
Only the image data detected in the set distance measuring range 110 is compared. The center of the distance measuring range 110 is set in advance.

If the image data do not match, the image data a on the left side of the photosensor array 12a is moved to the right and the image data b on the right side of the photosensor array 12b is moved to the left, as indicated by the broken line in FIG. After shifting and comparing, the shift amount when the image data a and b match each other is detected. Since the X-coordinates a _L1 and a _R1 of the image forming positions L ₁ and R ₁ match this shift amount, the distance detection circuit 14 shifts the shift amount a _L1 and
a The distance y from _R1 to the object is expressed by the above-mentioned equation 8
Can be calculated by However, since it is difficult to move the image data when calculating the distance data,
In reality, the distance measuring range 110 is moved to calculate the shift amount of the distance measuring range 110 when the image data match.

FIG. 8 is a diagram showing a line-shaped visual field of each sensor array when the object (intruder) 100 is viewed from the distance sensor 10 side. In FIG. 8, the optical sensor arrays 12a, 12c, 12 on the left side of the three pairs of optical sensor arrays are provided.
e has a plurality of sensor arrays arranged in rows. A measurement window composed of a plurality of consecutive sensor arrays is defined from the sensor array. And
For each measurement window, a sensor array representative of that measurement window is determined. In FIG. 8, the sensor array is located at the center of the measurement window. In this case, both end regions of the optical sensor array which cannot be the center are excluded from the target of the representative sensor array.

A measurement window number (hereinafter abbreviated as measurement window No.) is assigned corresponding to the arrangement order of the measurement windows thus determined. For example, the line No. corresponding to the visual field of the photosensor array 12e placed horizontally in FIG. 2, in order from the left measurement window to the right measurement window, 0, 1, 2, ...
Measurement window No. as shown in (total number of windows-1)
Turn on.

Here, the measurement window is set for each photosensor array of the upper line (line No. 2), the middle line (line No. 1), and the lower line (line No. 0) and on the photosensor array. It is defined to be arranged at equal intervals (WP). Then, the distance detection circuit 14 displays the measurement window No. For each measurement, the mutual correlation between the images of the objects in the pair of measurement windows is calculated, and the distance is calculated according to the above-mentioned principle. The distance data obtained for each is output.

Next, the principle of intruder determination using distance data will be described. 9, 10, 11, and 12 are flowcharts showing the operation procedure of the microcomputer 20. FIG. 9 is a flowchart regarding measurement after the distance sensor power is turned on. After the power is turned on in step S1, initialization is performed to set the intruder detection flag = 0 in step S2. This intruder detection flag indicates whether or not an intruder has been detected in the previous measurement. Intruder detection flag =
When it is 0, it means that the intruder was not detected in the previous measurement, and when the intruder detection flag = 1, it means that the intruder was detected in the previous measurement. Immediately after the power is turned on, the intruder detection flag is naturally set to 0.

Next, in step S3, distance measurement is performed based on the distance measurement principle described above. Distance measurements are made for all line windows. The latest distance data at the measurement point where the line is i and the window is j is represented by d _new (i, j). Next, in step S4, background distance data is created. Specifically, the process jumps to the background distance data creation subroutine shown in FIG.

Step S7 of the subroutine shown in FIG.
In, if the intruder is detected in the previous measurement, that is, if the intruder detection flag = 1, the process is terminated and the process returns to the main routine. At this time, the background distance data continues to hold the data up to the previous time. On the other hand, if no intruder has been detected in the previous measurement, that is, if it is determined that the intruder detection flag = 0, then in the next step S8, distance measurement is performed multiple times, that is, N times.

In step S9, the line number. Perform the initial setting of. That is, i = 0 is set. Next, in step S10, the window number. Perform the initial setting of. That is, j = 0 is set.

Next, in step S11, the background distance is set to d.
_{It is calculated as average} (i, j). The background distance data d _average (i, j) is an average value calculated below.

[0059]

[Equation 9]

Here, dt (i, j) has a line i,
The distance data at the t-th time out of a plurality of (N times) continuous measurements at the measurement point whose window is j is shown. Multiple times (N
It is possible to create background distance data that is less affected by noise by measuring the number of times and _obtaining the average value d _average (i, j). Further, since the background distance data is created only when no person is detected in the measurement, it is possible to create reliable background distance data when no person is present.

Then, the process proceeds to step S12. Step S1
If j is not the maximum window No. 2, 1 is added to j in step S14 to generate a new j, and then the process returns to state C and the same processing is repeated from the beginning of step S11. If j is the maximum window No in step S12, the process proceeds to step S13. If i is not the maximum line number in step S13, step S13
After adding 1 to i in 15 to generate a new i, the process returns to the state B, and the same processing is repeated from the beginning of step S10. In step S13, i is the maximum line number.
If it is determined that, the process exits the subroutine and returns to the main routine.

After returning to the main routine, the process proceeds to step S5 in FIG. In step S5, difference distance data is created. Specifically, the process is performed by jumping to the difference distance data creation subroutine shown in FIG. Figure 1
In step S16 of the subroutine shown by 1, the line No. is initialized. That is, i = 0 is set. Next, in step S17, the window No. is initialized. That is, j = 0 is set.

In step S18, the difference between the background distance data and the latest distance data is calculated. The difference distance data of the measurement point where the line is i and the window is j is d (i, j)
It is represented by. Further, if the difference between the background distance data and the latest distance data is represented by df, it can be represented by the following equation.

[0064]

## EQU10 ## Difference df = d _average (i, j) -d
_new (i, j)

Next, in step S19, the judgment distance and the difference d
If f is compared and df> determination distance is satisfied, that is, if it is determined that the distance data at the same location has changed, the process proceeds to step S20. If the distance data has not changed, the process jumps to step S21. When the process proceeds to step S20, the difference distance data d (i, j) has d _new.
When (i, j) is substituted and the process proceeds to step S21,
As the difference distance data, d (i, j) = invalid distance data (for example, data such as 0) is substituted.

As a result, the background distance becomes invalid distance data. Therefore, if no determination is made in the case of invalid distance data, an intruder can be detected without being affected by the background. Then, the process proceeds to step S22. In step S22, j
If it is determined that is not the maximum window, 1 is added to j in step S24 to generate a new j, the state E is returned to, and the same processing is repeated from the beginning of step S18. J is the maximum window number in step S22
If it is determined to match, the process proceeds to step S23.

If it is determined in step S23 that i is not the maximum line number, i is determined in step S25.
After adding 1 to generate a new i, return to state D,
The same process is repeated from the beginning of step S17. If it is determined in step S23 that i is the maximum line No, the process exits the subroutine and returns to the main routine.

After returning to the main routine, the process proceeds to step S6 in FIG. In step S6, an intruder is detected from the difference distance data. Specifically, it is performed by jumping to the intruder detection subroutine shown in FIG. In step S26 of the subroutine shown in FIG. 12,
The intruder detection flag is initialized to 0.

In step S27, the line No. is initialized. That is, i = 0 is set. Next, in step S28, the window No. is initialized. That is, j = 0 is set.

Next, in step S29, it is determined whether or not the difference distance data is invalid distance data. If the difference distance data is invalid distance data, the process jumps to step S35. On the other hand, if the difference distance data is not invalid distance data, the process proceeds to step S30 and the measurement target 100
The height of is calculated based on the following formula.

[0071]

H = H _s −d (i, j) / tan θ _i

Here, h is the detection height, H _s is the mounting height,
θ _i is the viewing direction angle of the line i with reference to the vertical direction (see FIG. 13).

FIG. 13 is a side view of the distance sensor 10 viewed from the horizontal direction A shown in FIG. Of these, m ₀ ,
m ₁ and m ₂ are distances from the optical axis of lines 0 to 2, respectively. The signs of m ₀ , m ₁ and m ₂ are positive in the upward direction and negative in the downward direction. For example, in FIG. 11, the signs of m ₀ and m ₁ are positive, and the sign of m ₂ is negative. The above calculation is performed assuming that the vertical direction of the wall 1a to which these distance sensors are attached coincides with the vertical direction.

If θ _s is the mounting angle, θ _i is line 0
2 to 2 are calculated as follows based on FIG.

[0075]

[Equation 12] θ ₀ = 90 ° − {tan ⁻¹ (m ₀ / f) + θ _s }

[0076]

[Equation 13] θ ₁ = 90 ° − {tan ⁻¹ (m ₁ / f) + θ _s }

[0077]

[Equation 14] θ ₂ = 90 ° − {tan ⁻¹ (m ₂ / f) + θ _s }

The heights are obtained by substituting these θ ₀ , θ ₁ , and θ ₂ into the _equation 11 described above. After calculating the height h of the object in step S30, the process proceeds to step S31.

In step S31, it is determined whether the height h is higher than a predetermined determination height. If it is determined in step S31 that h> determination height is satisfied, the object 100 is determined to be an intruder in step S32, and the intruder detection flag = 1 is set in step S34. On the other hand, if it is determined that h> judgment height is not satisfied, it is determined that the object 100 is not an intruder, and the process jumps to step S33. In step S33, it is determined that the measurement object is not an intruder.

An appropriate value (for example, 1 m, which is less than the height of most adults) is set as the above-mentioned determination height to determine whether the object 100 is a person or not. As a result, it is further determined in step S31, S32, S33, and S34 whether the object 100 determined to be a moving body is a person or a small animal, and erroneous detection is prevented.

Then, the process proceeds to step S35. Step S
If it is determined in step 35 that j is not the maximum window No, 1 is added to j in step S37 to obtain a new j.
After generating, the process returns to the beginning of step S29, and steps after step S29 are repeated. In step S35, j is the maximum window number. If it is determined that, the process proceeds to step S36.

In step S36, i is the maximum line N
o. If not, i is determined in step S38.
1 is added to generate a new i, and then step S28
Returning to the beginning, the steps after step S28 are repeated. On the other hand, in step S36, i is the maximum line N
o. If so, exit the subroutine and return to the main routine. In FIG. 9, the process proceeds from step S6 to state A, returns to the beginning of step S3, and repeats this operation thereafter to always detect an intruder.

As described above, according to the present embodiment, since the viewing angles are made different by the three photosensor arrays, as is apparent from FIGS. 4 and 5, where the intruder is located in the room. However, it is possible to reliably detect. For example, it becomes possible to detect the intruder shown by the dotted line in FIG. Further, in the case where the room 1 is narrow, at least one optical sensor array 12 can detect, and if the arrangement of the distance sensor 10 and the measurement visual field angle are appropriately adjusted, it can be sufficiently used. .

Further, in the present embodiment, three pairs of photosensor arrays are used, but the invention is not limited to three pairs, and one or two or four pairs of photosensor arrays may be used. . If the number of photosensor arrays is large, the detectable area is widened as is apparent from FIGS. 4 and 5, and therefore, it can be used even in a large room 1. In this way, the number of optical sensor arrays, the measurement viewing angle, and the mounting height are set appropriately according to the size of the room.

Further, in the present embodiment, the left and right two photosensor arrays are used as a pair, but for example, one photosensor array having a length of two may be treated as if there were virtually one photosensor array. . In the present invention, when referring to a pair of optical sensor arrays, such a long optical sensor array may be used as a pair of optical sensor arrays. Further, in the present embodiment, two left and right photosensor arrays are paired and a plurality of them are arranged in the vertical direction. However, for example, the sensor arrays are arranged in a matrix over a plurality of rows and a plurality of columns. Even if the photosensor array is on a flat surface like this, the window number and line number
Can be set, the distance can be calculated by the distance measuring means, and the present invention can be implemented. These are appropriately selected and designed in consideration of costs and the like. The first embodiment is such.

Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 17 is a schematic configuration diagram of the second embodiment of the present invention. The present embodiment is different from the first embodiment shown in FIG. 1 in that the microcomputer 20 ′ including the background distance data updating means 25 newly added in FIG.
It is an intruder detection device that is different in that it is equipped with. The background distance data updating means 25 is means for updating the background distance data output from the background distance data creating means 21. Since the principle of distance data calculation / distance measurement is the same as that of the first embodiment, duplicate description will be omitted.

Next, the background distance data creating means 21, the difference distance data creating means 22, the calculating means 23, the intruder determining means 24, and the background distance data updating means 25 will be described. 18 to 22 show the microcomputer 2
It is a flowchart which shows the operation procedure of 0 '. FIG.
3 is a flow chart of measurement after the power of the distance sensor 10 is turned on.

After the power is turned on in step S101, background distance data is created in step S102. Specifically, the process jumps to the background distance data creation subroutine shown in FIG. 19 is a flowchart of the background distance data creating means 21 in detail. First, step S107
In, the measurement is performed multiple times, that is, N times. In step S108, the line number. Perform the initial setting of.
That is, i = 0 is set. Next, in step S109, the window number. Perform the initial setting of. That is, j =
It is set to 0. Then determined in step S110 the background distance data as _{d av erage (i, j)} . The background distance data d _average (i, j) is an average value calculated by the following equation.

[0089]

[Equation 15]

Here, dt (i, j) has a line i,
The distance data at the t-th time out of a plurality of (N times) continuous measurements at the measurement point whose window is j is shown. Multiple times (N
It is possible to create background distance data that is less affected by noise by measuring the number of times and _obtaining the average value d _average (i, j).

Then, the process proceeds to step S111. Step S
In 111, j is the maximum window number. If not, after adding j in step S113, the state is returned to the state c, and the process is repeated from step S110. In step S111, j is the maximum window number. If so, step S112
Proceed to. In step S112, i is the maximum line N
o. If not, i is added in step S114, the state is returned to state b, and the process is repeated from step S109. In step S112, i is the maximum line number. If so, the process ends. Then, the process returns to step S102 of the main routine shown in FIG.

Then, the process proceeds to step S103. In step S103, distance measurement is performed based on the distance measurement principle described above. Distance measurements are made for all line windows. The latest distance data at the measurement point where the line is i and the window is j is represented by d _new (i, j). Then, it progresses to step S104.
Step S104 in FIG. 18 is the difference distance data creating means 2
The difference distance data is created in step S104. Specifically, the process jumps to the difference distance data creation subroutine shown in FIG. FIG. 20 is a flowchart of the difference distance data creating means 22 in detail. In step S115, the line number. Perform the initial setting of.
That is, i = 0 is set. Next, in step S116, the window number. Perform the initial setting of. That is, j =
It is set to 0.

In step S117, the difference between the background distance data and the latest distance data is calculated. The difference distance data at the measurement point where the line is i and the window is j is d (i, j)
It is represented by. Further, when the difference between the background distance data and the latest distance data is represented by df, the following equation is obtained.

[0094]

## EQU16 ## Difference df = d _average (i, j) -d
_new (i, j)

Next, in step S118, if the difference df> judgment distance, the process proceeds to measurement step S119, and if not, the process proceeds to step S120. Step S119
When the process proceeds to, the differential distance data is expressed by the following equation.

[0096]

D (i, j) = d _new (i, j)

When the process proceeds to step S120, the difference distance data is expressed by the following equation.

[0098]

D (i, j) = invalid distance data

As a result, since the distance data only for the background becomes invalid distance data, it is possible to detect an intruder that is not affected by the background. Then, the process proceeds to step S121. In step S121, j is the maximum window number. If not,
After adding j in step S123, the state returns to state e,
Repeat from step S117. In step S121, j is the maximum window number. If so, step S12
Go to 2. In step S122, i is the maximum line number. If not, i is added in step S124, the state is returned to the state d, and the process is repeated from step S116.
In step S122, i is the maximum line number. If so, the process is terminated, and the process returns to step S104 in FIG.

Then, the process proceeds to step S105. The step S105 is a step of detecting the intruder detection from the difference distance data, and the calculating means 23 and the intruder judging means 24.
Equivalent to. Specifically, the process jumps to a subroutine for detecting intruder detection from the difference distance data shown in FIG. FIG. 21 is a flowchart of means for detecting intruder detection from the difference distance data. First, the calculation means will be described. In step S125, the line number. Perform the initial setting of. That is, i = 0 is set. Next, in step S126, the window number. Perform the initial setting of. That is, j = 0 is set.

Next, in step S127, if the difference distance data is invalid distance data, the process proceeds to step S132, and if the difference distance data is not invalid distance data, step S132.
Proceed to 128. In step S128, the height of the object 100 is calculated based on the following formula.

[0102]

H = H _s −d (i, j) / tan θ _i

Here, h: detected height H _s : mounting height θ _i : viewing direction angle of line i with reference to the vertical direction (see FIG. 13). Where θ _i is line 0
2 to 2 are calculated as follows based on FIG.

[0104]

[Equation 20] θ ₀ = 90 ° − {tan ⁻¹ (m ₀ / f) + θs}

[0105]

[Equation 21] θ ₁ = 90 ° − {tan ⁻¹ (m ₁ / f) + θs}

[0106]

[Equation 22] θ ₂ = 90 ° − {tan ⁻¹ (m ₂ / f) + θs}

However, θs is a mounting angle.

FIG. 13 is a view of the distance sensor 2 viewed from the horizontal direction A in FIG. 1, and also m ₀ , m ₁ , m ₂
Are the distances from the optical axis direction of lines 0 to 2, respectively.
The signs of m ₀ , m ₁ and m ₂ are positive in the upward direction and negative in the downward direction. For example, in FIG. 13, the signs of m ₀ and m ₁ are positive,
The sign of m ₂ is negative. Further, in FIG. 13, the above calculation is performed on the assumption that the vertical direction of the wall 1a to which the sensor is attached coincides with the vertical direction. In this way, the height h of the object 100 is calculated in step S128, and then the process proceeds to step S129.

In step S129, the intruder determination means 24
And if the above height h> determined height, then step S1
In step 30, the object 100 is determined to be an intruder. If not, the process proceeds to step S131, and it is determined that the target object 100 is not an intruder. The above judgment height is the object 10
An appropriate value is set to determine if 0 is a person or not. This prevents erroneous detection when the object 100 is a small animal.

Then, the process proceeds to step S132. Step S
132, j is the maximum window number. If not, after j is added in step S134, step S127
Returning to step S127, steps S127 and thereafter are repeated. Step S
132, j is the maximum window number. If so, the process proceeds to step S133. In step S133, i
Is the maximum line number. If not, after i is added in step S135, steps S126 and thereafter are repeated.
In step S133, i is the maximum line number. If so, the process is terminated and the process returns to the main routine.

Then, the process proceeds to step S106 in FIG. Step S106 is background distance data updating means 25.
Equivalent to. Specifically, the process jumps to the background distance data update subroutine shown in FIG. FIG. 22 is a flowchart of the background distance data updating means in detail. The background distance data is updated at a preset cycle. Here, the background distance data is updated once every M times of the number of measurements.

First, in step S136, if the number of measurements is a multiple of M, the process proceeds to step S137. If the number of measurements is not a multiple of M, the background distance data updating process is terminated and the process returns to the main routine. In step S137, the line number. Perform the initial setting of. That is, i = 0 is set. Next, in step S138, the window N
o. Perform the initial setting of. That is, j = 0 is set.
Next, in step S139, the current background distance data d
_{The average} is updated by the moving average process based on the following equation using the latest distance data d _new .

[0113]

D _average (i, j) = d
_average (i, j) × (1-1 / M) + d _new
/ M

As a result, the background distance data is gradually updated at a cycle of once every M times, so that an intruder can be detected without being affected by changes in the background object. The value of M is decreased in an environment where the change of the background object is large, and conversely, the value of M is increased in the environment where the change of the background object is small.

Then, the process proceeds to step S140. Step S
In 140, j is the maximum window number. If not, after j is added in step S142, the state is returned to step h and the process is repeated from step S139. In step S140, j is the maximum window number. If so, step S141
Proceed to. In step S141, i is the maximum line N
o. If not, after adding i in step S143, the process returns to state i and repeats from step S138. In step S141, i is the maximum line number. If so, the background distance data update process is terminated, and the process returns to the main routine. In the main routine in FIG. 18, the state a
By returning to the above and repeating this operation, the intruder is always detected.

In the second embodiment, since the background distance data is newly updated, the background distance data is not affected by a change in the background object (for example, a change in daily life or daily work such as a flower in the room). Intruders can be detected.

[0117]

As described above, according to the invention described in claim 1, image data of an object is obtained by using the distance sensor and the optical sensor array, and the signal processing unit / distance detection circuit detects the object. By detecting the distance data, the height of the object is calculated from this distance data, the mounting height of the distance sensor, and the tilt angle that is the intersection angle between the visual field direction of each optical sensor array and the vertical direction. , The object is highly reliable because it does not accidentally detect a small animal as an intruder in order to determine that this object is an intruder when it has the height of a human and moves. It is possible to provide an intruder detection device.

Further, in this intruder detecting device, the distance sensor does not project the light beam to the intruder, but the intruder is compared by comparing the images on the optical sensor array obtained by the reflected light of the natural light from the intruder. In order to obtain the distance to the intruder, it is possible to obtain the distance to the intruder without being influenced by the distance to the intruder or the reflectance of the intruder, and the intruder can be reliably detected. Furthermore, since the background information is removed and only the moving body is detected, it is possible to reliably prevent the situation where the intruder is erroneously detected from the background information.

According to the second aspect of the present invention, the intruder detection device is an intruder determination means for determining that the object is an intruder when the height of the object is higher than the determination height. Since it is provided, it is possible to detect an intruder with high accuracy without determining a small animal that is shorter than a person as an intruder.

Further, according to the invention described in claim 3, in this intruder detecting device, the distance sensor is provided with a plurality of pairs of optical sensor arrays. As a result, it is possible to reliably detect an intruder regardless of the location of the intruder, even in a wide area such as a room or facility. Furthermore, it is possible to prevent a situation where the intruder cannot detect because the intruder jumps over or slips through the detection area.

Further, according to the invention described in claim 4, the image signal output from one window defined by a part of a plurality of arrays among all the arrays divided into a plurality of columns is used. As a result, the shift amount can be calculated by the program processing.

Further, according to the invention of claim 5, since the average value of the distance data measured a plurality of times is used as the background distance data, it is possible to generate the background distance data less affected by noise.

Further, according to the invention of claim 6, since the background distance data is created when the intruder is not detected in the previous measurement, the reliable background distance data not including the information of the intruder is created. It is possible to

Further, according to the invention described in claim 7, in the intruder detecting device, in the intruder determining means, only for distance data in which the difference obtained by subtracting the latest distance data from the previous distance data is larger than the determination distance, The data was valid, and the rest was uniformly invalid distance data. This makes it possible to exclude all background information such as furniture and equipment of stationary objects from the target of intruder determination, and use only data that is a moving body such as an intruder.

Further, according to the invention of claim 8, in the intruder detecting device, in the intruder determining means, only the distance data in which the difference obtained by subtracting the latest distance data from the previous distance data is larger than the determination distance, It is determined whether the object is an intruder. That is, the stationary object is excluded from the object of the intruder determination as furniture and equipment, and only the moving body such as the intruder is determined. This enables highly accurate intruder detection.

Further, according to the invention of claim 9, in the intruder detecting device, the background distance data updating and setting means gradually updates the background, thereby enabling accurate detection even when the background changes. is there.

In general, according to the present invention, it is possible to provide an intruder detecting device for accurately detecting an intruder.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a first embodiment of the present invention.

FIG. 2 is a detailed configuration diagram of the distance sensor according to the first embodiment of the present invention.

FIG. 3 is a layout view of an optical sensor array.

FIG. 4 is a perspective view illustrating installation of an intruder detection device installed indoors.

FIG. 5 is a side view illustrating installation of an intruder detection device installed indoors.

FIG. 6 is a diagram illustrating the principle of distance detection.

FIG. 7 is an operation explanatory diagram of a distance detection circuit.

FIG. 8 is a diagram showing a linear field of view of each optical sensor array when an object (intruder) is viewed from the distance sensor side.

FIG. 9 is a flowchart showing an operation procedure of the microcomputer 20.

10 is a flowchart showing an operation procedure of the microcomputer 20. FIG.

11 is a flowchart showing an operation procedure of the microcomputer 20. FIG.

FIG. 12 is a flowchart showing an operation procedure of the microcomputer 20.

FIG. 13 is a side view of the distance sensor.

FIG. 14 is an explanatory diagram of a distance measuring principle.

FIG. 15 is an explanatory diagram of a correlation calculation based on the principle of distance measurement.

FIG. 16 is an explanatory diagram of a multi-point distance measuring principle.

FIG. 17 is a schematic configuration diagram of a second embodiment of the present invention.

FIG. 18 is a flowchart showing an operation procedure of the microcomputer 20 ′.

FIG. 19 is a flowchart showing an operation procedure of the microcomputer 20 ′.

FIG. 20 is a flowchart showing an operation procedure of the microcomputer 20 ′.

FIG. 21 is a flowchart showing an operation procedure of the microcomputer 20 ′.

FIG. 22 is a flowchart showing an operation procedure of the microcomputer 20 ′.

[Explanation of symbols]

10 distance sensor 11 imaging lens 11a, 11b imaging lens 12 optical sensor array 12a, 12b, 12c optical sensor array 12d, 12e, 12f optical sensor array 13 signal processing unit 14 distance detection circuit 15a, 15b amplifier 16a, 16b A / D converter 17 Storage device 20 Microcomputer 21 Background distance data creating means 22 Difference distance data creating means 23 Calculating means 24 Intruder judging means 25 Background distance data updating means 1 Indoor 1a Wall 3 Upper line (line No0) measurement field of view 4 Measurement field of the middle line (line No1) 5 Measurement field of the lower line (line No2) 100 Target (intruder) 110 Distance measuring range

Front page continuation (51) Int.Cl. ⁷ identification code FI theme code (reference) G06T 7/60 150 G06T 7/60 150J 180 180B // G01V 8/20 G01V 9/04 QF term (reference) 2F065 AA06 BB05 CC16 FF01 FF09 JJ02 JJ05 JJ25 LL05 QQ03 QQ31 5B057 AA19 BA02 BA13 CA11 CA16 DA12 DB01 DC03 DC36 5C084 AA02 AA07 AA14 BB05 BB31 CC17 DD12 EE02 GG43 GG52 GG54 GG78 5L096 FA02 CA03 CA02

Claims (9)

[Claims] 1. A pair of imaging lenses, which have parallel optical axes and are arranged to form images of objects on the same imaging plane, and arranged on the imaging planes of the pair of imaging lenses. An optical sensor array that outputs an image signal from an array at a position corresponding to the image formation by each of the pair of imaging lenses, and a distance to an object in the image is measured based on the image signal output from the optical sensor array. Distance detecting means for generating and outputting distance data by distance, and background distance data generating means for generating and outputting background distance data which is the distance to the background of the object based on the distance data output from the distance detecting means. A difference distance data creating unit that obtains a difference between the background distance data and the latest distance data, and outputs the latest distance data as difference distance data when the difference value exceeds a predetermined value, The difference distance data output from the split distance data creating means, the view direction angle that is the intersection angle between the view direction of the optical sensor array and the vertical direction with respect to the ground, and the mounting height of the optical sensor array are used to respectively detect An intruder determining means for determining whether or not the object is an intruder based on the detected height calculated by the calculating means; Detection device. 2. The intruder detection device according to claim 1, wherein the intruder determination means determines that the object is an intruder when the detected height of the object is higher than a predetermined determination height. An intruder detection device characterized by determining that 3. The intruder detection device according to claim 1, wherein a plurality of the photosensor arrays are arranged on the image plane with different heights so that the viewing direction angles are different. A pair of optical sensor arrays, wherein the plurality of pairs of optical sensor arrays detect image signals of an object with different viewing direction angles, and the distance detecting means, the background distance data creating means, the differential distance data creating means, An intruder detection device, characterized in that the calculating means and the intruder determining means determine intruders in different visual field directions. 4. The intruder detection device according to claim 1, wherein the optical sensor array is an array divided into a plurality of rows, and a part of the entire array is provided. 1 defined with multiple arrays
An intruder detection device characterized by using an image signal output from the measurement window. 5. The intruder detection device according to claim 1, wherein the background distance data creating means uses the average value of the distance data measured a plurality of times as the background distance data. Intruder detection device characterized by. 6. The intruder detection device according to claim 5, wherein the background distance data creating means uses distance data obtained a plurality of times after the intruder is not detected. apparatus. 7. The intruder detection device according to claim 1, wherein the difference distance data creating unit includes a background for each distance data for each measurement window on each sensor line. If the difference between the distance data and the latest distance data exceeds the judgment distance, substitute the latest distance data into the difference distance data,
An intruder detection device, which is means for substituting invalid distance data into the difference distance data when the difference between the background distance data and the latest distance data is less than or equal to the determination distance. 8. The intruder detection device according to claim 7, wherein the intruder determination means omits the determination when the previous difference distance data is invalid distance data, and determines the latest distance data. An intruder detection device, which is a means for performing. 9. The intruder detection device according to claim 1, further comprising background distance data updating means for updating the background distance data output from the background distance data creating means. Intruder detection device characterized by.