JPH06214014A - Approaching object detecting device - Google Patents

Abstract

PURPOSE:To enable the automatic, positive judgment of an approaching object by preparing a central value pattern on the basis of the feature quantity of each window area formed in an image, and computing this pattern relatively in time series. CONSTITUTION:An image G after being converted into luminance information by picture elements at an image processing part 22 is displayed on a display device 23 and inputted into a window processing part 24. The processing part 24 sets window areas W1-Wn, and displays them on the image G of the display device 23 and inputs them into a central value sampling part 25. The sampling part 25 extracts the feature quantity of the respective areas M1-Wn at the time t0, that is, the central value F1-Fn (brightness). A pattern preparing part 26 prepares a pattern P from the central value F1-Fn and inputs it into a storage part 27 and a shift quantity computing part 28. The pattern preparing part 26 also prepares a pattern P to the image G at the time t0+DELTAt after the lapse of micro time DELTAt, and the computing part 28 compares the latest pattern P with the preceding pattern P0 every DELTAt time so as to obtain shift quantity DELTAP. A judging part 29 judges the separating/approaching of a rear vehicle by the positive/negative of the shift quantity DELTAP.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an approaching object detecting device which is mounted on, for example, an automobile and is used to monitor the surroundings of the vehicle. The present invention relates to an approaching object detection device suitable for preventing accidents.

[0002]

2. Description of the Related Art Generally, when a vehicle running on a road is approached by another vehicle from the rear, the driver of the vehicle in front is less likely to notice the vehicle, which easily leads to an unexpected accident. In particular, when a running vehicle changes lanes, if there is a following vehicle approaching the changed lane, a collision may occur depending on the relative speed of the host vehicle. Therefore, in recent years, various devices have been proposed to notify the driver of the approach of a rear vehicle.

For example, as an approaching object detecting device for detecting a following approaching vehicle with respect to a traveling vehicle, a device for measuring the approaching speed by a Doppler radar using ultrasonic waves or radio waves, a vehicle in a specific area using the radar A device for determining
Alternatively, there is a device in which a marker is added to an image displayed by using an image pickup means such as a video camera so that the position of a rear vehicle can be confirmed.

FIG. 10 is a block diagram showing a conventional approaching object detection device disclosed in, for example, Japanese Patent Laid-Open No. 1-231600, showing an example of a Doppler radar using ultrasonic waves. In the figure, X is a Doppler radar that is provided at the rear of the vehicle and has ultrasonic wave transmission and reception functions, Y is an approach detection device that detects the approach of a vehicle behind based on the voltage signal from the Doppler radar X, and Z is the approach. The display device is driven by an output signal from the detection device Y.

The Doppler radar X includes a crystal oscillator 1 for generating an oscillation signal, a frequency divider 2 for dividing the oscillation signal, a transmission amplifier 3 for generating a transmission signal based on the frequency division signal,
A wave-transmitting piezoelectric vibrator 4 that is driven by a transmission signal to transmit an ultrasonic wave; and an ultrasonic wave transmitting unit 4a of the wave-transmitting piezoelectric vibrator 4,
A piezoelectric resonator 5 for receiving waves that receives reflected ultrasonic waves from a rear vehicle, an ultrasonic wave receiving portion 5a of the piezoelectric vibrator 5 for receiving waves, and a reception amplifier that amplifies a reception signal from the piezoelectric vibrator 5 for receiving waves. 6
And a frequency conversion circuit 7 for converting the frequency of the received signal based on the divided signal, and a frequency-voltage conversion circuit 8 for converting the converted frequency into a voltage signal.

The approach detection device Y includes a voltage comparison circuit 9 for comparing the voltage signal from the Doppler radar X with a predetermined voltage,
The output circuit 10 generates a drive signal for the display device Z based on the comparison result. The display device Z is composed of, for example, an LED array 11 as illustrated.

FIG. 11 is an explanatory view schematically showing a plurality of running vehicles. A is a vehicle traveling in front, B is a vehicle traveling in the rear, Va is a traveling speed of the front vehicle A, and Vb is a rear. The traveling speed of the vehicle B, fo is the frequency of the ultrasonic waves transmitted from the vehicle A in front.

Next, the operation of the conventional approaching object detection device shown in FIG. 10 will be described with reference to FIG. When the transmission amplifier 3 in the Doppler radar X drives the piezoelectric vibrator 4 for wave transmission and transmits an ultrasonic wave of frequency fo from the front vehicle A to the rear vehicle B, the ultrasonic waves reflected by the rear vehicle B are It is received by the piezoelectric resonator 5 for receiving waves.

At this time, a Doppler shift proportional to the approaching speed of the two vehicles A and B, that is, the relative speed ΔV (= Vb-Va) occurs. This Doppler shift is extracted by the frequency conversion circuit 7 as a beat between the reception signal and the transmission signal passed through the reception amplifier 6, and the beat frequency is converted into a voltage signal by the frequency-voltage conversion circuit 8 and output. .

The voltage signal representing the approach speed of the rear vehicle B with respect to the front vehicle A is compared with the reference voltage by the voltage comparison circuit 9 in the approach detection device Y and displayed on the display device Z as a signal representing the speed.

FIG. 12 is an explanatory view showing a conventional approaching object detection device described in, for example, Japanese Patent Publication No. 62-29265, which shows an example using a radar antenna. In this case, a radar antenna X'is provided at the side mirror position of the front vehicle A, a signal received by the radar antenna X'is displayed on a display device (not shown) in the vehicle A, the presence or absence of the rear vehicle B and the distance To detect.

However, the Doppler radar X shown in FIG.
Also, since the radar antenna X'of FIG. 12 radiates ultrasonic waves and radio waves, there is a risk that it will interfere with ultrasonic waves and radio waves from other vehicles having the same function, making analysis difficult. Further, when there are a plurality of objects, it is difficult to clearly detect the reflected signal because it is impossible to determine which object the reflected signal is from.

FIG. 13 is an explanatory view showing a display screen of a conventional approaching object detection device described in, for example, Japanese Patent Publication No. 3-47213, and shows an example using a video camera.
In this case, the image a captured by the video camera provided at the rear part of the vehicle is displayed on the display device inside the vehicle.
At this time, as shown in the figure, if a marker b indicating the inter-vehicle distance is added in the screen, the approximate distance to the rear vehicle B and the approaching state can be recognized. However, when the image a is displayed in this way, the driver has to keep an eye on the image a at all times even though the driver is driving.

[0014]

As described above, the conventional approaching object detection device uses the Doppler radar X (FIG. 10) or the radar antenna X '(FIG. 12) to detect the distance to the rear vehicle B and the relative speed. Or the position of the rear vehicle B is recognized using the image a (FIG. 13) from the video camera and the marker b.

However, when the Doppler radar X or the radar antenna X'is used, the direction and position of the rear vehicle B cannot be detected and the signals interfere with each other, so that a plurality of objects exist. However, there is a problem that the rear vehicle B cannot be specified and detected.

When the image a obtained by the video camera is used, the device itself does not have a determination function and merely displays the image a. Therefore, the driver can determine the distance to the rear vehicle B and the relative speed. You must judge empirically. Therefore, there is a problem that it is not possible to reliably determine the presence or absence of danger, and the driver may not concentrate on driving the own vehicle, which may be dangerous.

The present invention has been made in order to solve the above problems, and an object thereof is to provide an approaching object detection device capable of automatically and reliably determining an approaching object.

[0018]

According to a first aspect of the present invention, there is provided an approaching object detection device, which comprises an image pickup means for picking up an image,
A window processing unit that divides an image into a plurality of window regions, a representative value extraction unit that extracts the characteristic amount of each window region as a representative value, and a pattern creation unit that creates a pattern corresponding to the image using each representative value as a data string. And a shift amount calculation unit that compares the patterns of images at different imaging timings to calculate a shift amount, and a determination unit that generates a determination signal when the shift amount is greater than or equal to a predetermined value indicating the presence of an approaching object. It is equipped with.

According to a second aspect of the present invention, there is provided an approaching object detection device according to the first aspect, wherein the image pickup means is installed in the vehicle to capture a rear image of the vehicle.

According to a third aspect of the present invention, there is provided the approaching object detection device according to the second aspect, wherein the controller generates an alarm drive signal in response to the determination signal, and an alarm means driven by the alarm drive signal. Is provided.

According to a fourth aspect of the present invention, there is provided the approaching object detection device according to the third aspect, further comprising a switch means that is operated when the traveling direction of the vehicle is changed, and the image pickup means corresponds to the traveling direction of the vehicle. An alarm is issued when a plurality of rear images are captured, the determination unit generates a determination signal for each rear image, and the controller responds to the operation signal from the switch means and detects the determination signal for the changed traveling direction. A drive signal is generated.

According to a fifth aspect of the present invention, there is provided an approaching object detection device according to any one of the second to fourth aspects.
The window processing unit includes a turning detection unit that detects a turning state of the vehicle, and the window processing unit changes the set position of the window region in response to a turning signal from the turning detection unit.

According to a sixth aspect of the present invention, there is provided an approaching object detection device according to any one of the first to fifth aspects.
The window area is set along the radiation from the gazing point at infinity, and the pattern is composed of a one-dimensional data string.

According to a seventh aspect of the present invention, there is provided an approaching object detection device according to any one of the first to fifth aspects.
The window region is set along the radiation from the gazing point at infinity and is set along the direction perpendicular to the radiation, and the pattern is composed of a two-dimensional data string.

[0025]

According to the first aspect of the present invention, the representative value pattern is created based on the feature amount of each window region formed in the image, and the representative value pattern is calculated in time series to perform the approaching operation. Automatically and reliably determine the target object.

According to the second aspect of the present invention, the rear target object approaching the vehicle is automatically and surely determined.

Further, according to the third aspect of the present invention, when the object approaching rearward is judged, the alarm means is driven to alert the driver to ensure safety.

Further, according to claim 4 of the present invention, when it is determined that there is a rear object approaching on the changed traveling road before the traveling direction is changed, the alarm means is driven to warn the driver. Encourage and ensure safety.

Further, according to the fifth aspect of the present invention, the wind area is optimally updated and set in response to the turning direction of the vehicle, and the rear object approaching on the traveling path after the turning is surely determined.

Further, in the sixth aspect of the present invention, the window region along the radiation from infinity is set, and the approaching object is determined based on the one-dimensional pattern.

According to a seventh aspect of the present invention, a window region is set along the radiation from infinity and in a direction perpendicular to the radiation, and the approaching object is trusted based on the two-dimensional pattern. Make a good judgment.

[0032]

【Example】

Example 1. A first embodiment of the present invention (corresponding to claim 1, claim 2 and claim 6) will be described below with reference to the drawings. Figure 1
2 is a block diagram showing a functional configuration of Embodiment 1 of the present invention, and 20 is an approaching object detection means composed of the following 21 to 29.

Reference numeral 21 denotes an image pickup means or camera which is installed on the door mirrors or the like on both sides of the vehicle and takes rear and side images G, and 22 an image processing unit which processes the image G as data.
Reference numeral 23 is a display device for displaying the data-processed image G, and 24 is further data processing for the image G, and a plurality (n) of window regions
W1 to Wn are divided into window processing units, 25 is a representative value extraction unit that extracts the characteristic amount of each region as representative values F1 to Fn, and 26 is a pattern P corresponding to the image G by patterning the respective representative values F1 to Fn. Is a pattern creation unit for creating.

A storage unit 27 stores the pattern P delayed by a predetermined time interval Δt and stores it as a previous pattern Po, and a pattern 28 of each image at different imaging timings, that is, a storage unit.
A shift amount calculator that compares the previous pattern Po and the current pattern P in 25 to calculate the pattern shift amount ΔP, 29
Is a determination unit that generates a determination signal D indicating the presence of a rear vehicle approaching the host vehicle when the shift amount ΔP is equal to or greater than a predetermined value. The predetermined value is preset to a level indicating the presence of an approaching object (subsequent approaching vehicle).

Output signal of window processing unit 24 and determination unit 29
The determination signal D from is input to the display device 23 as necessary, and the states of the window regions W1 to Wn and the determination result are displayed on the display screen.

FIG. 2 is an explanatory view showing an example of the rear image G photographed by the camera 21 and displayed on the display device 23.
W7 is, for example, a window area divided into seven, and B is a rear vehicle photographed so as to enter the window areas W1 to W7. Here, as the image G, a case is shown in which the rear side on the lane side adjacent to the right of the traffic lane of the vehicle is photographed.
Further, the window areas W1 to W7 are set along the traveling path along the radiation from the gazing point at infinity.

FIG. 3 is an explanatory diagram showing the previous pattern Po and the current pattern P formed on the basis of the representative values F1 to F7 of the window regions W1 to W7 in FIG. 2, and (a) is a certain time t.
the previous pattern Po consisting of the representative values f1 to f7 at o,
(b) is a current pattern P including representative values F1 to F7 at time to + Δt after Δt (for example, 0.1 seconds) has elapsed.

FIG. 4 is a characteristic diagram showing the relationship of the representative value deviation ΔF with respect to the shift correction amount ΔPf of the pattern P, which is used in the processing within the shift amount calculating section 28. Here, the shift correction amount ΔPf is an integer, and corresponds to, for example, the amount by which the current pattern P in FIG. 3B is shifted in the decreasing direction of the window region number, and the representative value deviation ΔF is the respective window regions W1 to Wn. It corresponds to the sum of the deviations between the previous pattern P ′ and the current pattern P for each time. Further, the shift correction amount ΔPf that minimizes the representative value deviation ΔF is obtained as the shift amount ΔP.

Next, the operation of the first embodiment of the present invention shown in FIG. 1 will be described with reference to FIGS. The image G captured by the camera 21 is converted into luminance information for each pixel by the image processing unit 22, and then displayed on the display device 23 and input to the window processing unit 24.

The window processing unit 24 includes n window regions.
W1 to Wn, for example, seven window areas W1 to W7 as shown in FIG.
Are set, are input to the display device 23 to be displayed on the image G, and are input to the representative value extraction unit 25. The window regions W1 to Wn are set along the radiation road from the gazing point at infinity on the traveling lane, as described above.

The representative value extraction unit 25 sums the brightness of all the pixels in each of the window regions W1 to Wn as the feature amount of each of the window regions W1 to Wn, that is, the representative values F1 to Fn at a certain time to. The extracted value is extracted.

In addition, the pattern creating section 26 uses the respective window areas.
Representative values F1 to Fn for each of W1 to Wn as a one-dimensional data string are shown in FIG.
The pattern is created as shown in FIG.
And the shift amount calculator 28. This pattern P is
In the shift amount calculation unit 28, the current pattern P having the representative values F1 to Fn is compared with the previous pattern Po, and the storage unit
In 27, it is stored as the previous pattern Po composed of the representative values f1 to fn.

Similarly, the window region W1 is set for the image G at time to + Δt, which has passed for a minute time Δt (about 0.1 second).
A pattern P composed of representative values F1 to Fn for each Wn is created,
It is input to the storage unit 27 and the shift amount calculation unit 28. continue,
The shift amount calculation unit 28 compares the current pattern P with the previous pattern Po in the storage unit 27 for each minute time Δt, and performs a calculation for obtaining the shift amount ΔP.

That is, the pattern P of different imaging timing
For each of the representative values of Po and Po, the representative value deviation ΔF is successively calculated while changing the shift correction amount ΔPf for the pattern P this time, and the shift correction amount ΔPf that minimizes the representative value deviation ΔF is matched with the patterns Po and P. It is calculated as the shift amount ΔP for the purpose. The calculation of each representative value deviation ΔF is
It is represented by the summation formula (1) with k = 1 to n as follows.

ΔF = (1 / n) Σ | f (k) −F (k + ΔPf) | (1)

In equation (1), f (k) is the previous pattern P
A representative value of the k-th window region Wk of o, F (k + ΔP
f) is the (k + ΔPf) th representative value of the pattern P this time. Therefore, the representative value deviation ΔF is the previous pattern Po when the pattern P is shifted this time by the shift correction amount ΔPf.
Represents a value obtained by normalizing the total sum of the deviations of the representative values for each window region.

Thus, by repeatedly calculating the equation (1) while changing the shift correction amount ΔPf, the characteristic diagram of FIG. 4 is obtained, and the shift amount Δ that minimizes the representative value deviation ΔF is obtained.
P is required. This shift amount ΔP is equal to each pattern Po.
And P are values that best match each other.

Subsequently, the determination unit 29 determines that the rear vehicle B is a vehicle that is separating from the own vehicle when the shift amount ΔP is a positive value, and the rear vehicle B is the own vehicle when the shift amount ΔP is negative. It is determined that the vehicle is approaching. If the shift amount ΔP is equal to or greater than a predetermined value and the approaching rear vehicle B is detected, the determination signal D indicating the presence of the approaching object is generated. The determination signal D is used to notify the driver of the danger as described later.

The image G in FIG. 2 will be described in detail. Assuming that the reflected light on the road surface is small, the previous pattern Po at time to is as shown in FIG.
Typical value f3 near the window regions W3 and W4 corresponding to the position of
And f4 are maximum. Assuming that the rear vehicle B is approaching the host vehicle, the current pattern P at time to + Δt is the front window in the image G as shown in FIG.
The representative values F2 and F3 around W2 and W3 are maximum.

At this time, the shift amount ΔP is almost "-1".
Therefore, if the absolute value of the shift amount ΔP is compared with a predetermined value (for example, 1), | ΔP | ≧ 1. Therefore, the determination unit 29
Generates a determination signal D indicating that the rear vehicle B is approaching because the shift amount ΔP is equal to or greater than a predetermined value.

If the rear vehicle B does not exist, each of the window regions W1 to Wn exclusively displays the road surface. Therefore, even if the patterns Po and P are compared while changing the shift correction amount ΔPf, the representative value deviation is obtained. The minimum point of ΔF (see FIG. 4) does not occur. Therefore, a non-existent approaching object will not be erroneously detected.

When a shadow or the like exists on the road surface,
Although the movement of the shadow is detected by the change of the pattern P, since the own vehicle is traveling, the minimum point position of the representative value deviation ΔF, that is, the shift amount ΔP becomes positive, and is not erroneously detected as an approaching object.

In this way, a plurality of windows W1 to Wn are set in the image G from the camera 21, and the representative value of each window area is set.
By creating a pattern P based on (brightness) and obtaining the shift amount ΔP based on the time-series correlation processing of the pattern P, the rear vehicle B approaching, the object being separated, etc. are automatically and accurately Can be determined.

Further, since the image G of the camera 21 is automatically analyzed, signals such as ultrasonic waves and radio waves are not radiated, so that erroneous detection due to interference or interference does not occur. Further, even if there are a plurality of objects, the minimum value of the representative value deviation ΔF occurs at a plurality of locations, so that the approaching state of each object can be determined.

Example 2. In the first embodiment, the case where the approaching object detecting means 20 is installed in the vehicle to detect the approaching rear vehicle B has been described as an example, but it is necessary to detect the approaching object. Of course, it can be applied in other fields. For example, it can be used to detect and determine an approaching object at an intersection, and the determination signal D when the approaching object is detected is not limited to displaying or warning, but can be used arbitrarily.

Example 3. Further, the approaching object detection means 20 is configured to simply generate the determination signal D. However, when the approaching rear vehicle B is determined, the determination result by the determination signal D is displayed on the display device 23 as necessary. You may display an alarm. For example, a red marker or the like can be added to the rear vehicle B in the display image G so that it can be easily noticed, and the marker can be made to blink.

Example 4. (Corresponding to claim 3) Further, although not shown, in addition to the display device 23, an alarm means capable of giving an alarm by voice or the like, and a controller for generating a drive signal of the alarm means in response to the determination signal D are provided, and are approaching. The warning means may be driven in response to the determination signal D when the vehicle B behind the vehicle indicates a very dangerous state. This makes it possible to reliably notify the driver of a particularly dangerous approaching object.

Example 5. (Corresponding to Claim 4) In the fourth embodiment, the warning drive signal is generated only based on the determination signal D without considering the driving state of the host vehicle. An alarm drive signal may be generated in response to the change of direction. FIG. 5 is a block diagram showing a fifth embodiment of the present invention, and 20 is the same as that described above.

Reference numeral 31 is a switch means operated when the traveling direction of the host vehicle is changed, that is, a winker switch, 32 is a controller for judging the presence or absence of the judgment signal D in response to an operation signal S from the winker switch 31, and 33 is a controller 32. The alarm means is driven by the alarm drive signal E from.

In this case, the determination signal D includes a plurality of determination results for each image G in different directions. In addition, the controller 32 responds to the operation signal S to selectively detect the determination signal D for the changed traveling direction, and when the determination signal D is detected, generates the alarm drive signal E. .

For example, when the driver intends to change the course to the overtaking lane on the right side in the traveling direction of the host vehicle, when the driver operates the winker switch 31 to instruct the traveling direction change, the operation signal S indicating a right turn is made. In response to this, the controller 32 detects the presence or absence of the determination signal D based on the right rear image.

At this time, if an approaching object, that is, the rear vehicle B (see FIG. 2) is present in the right lane, an alarm drive signal E is generated in response to the determination signal D, and the alarm means 33 is driven. Call the driver's attention by voice. Therefore, the driver
When the blinker switch 31 is operated, it is possible to immediately stop turning right by recognizing that there is an approaching rear vehicle B on the lane after the change, and prevent a particularly dangerous accident when changing the lane. it can. Further, in order to prevent the accident more reliably, it may be made difficult to turn the steering wheel to the right in response to the determination signal D.

Example 6. (Corresponding to claim 5) Further, the predetermined window area is set on the assumption that the vehicle is traveling straight, but the window area may be updated and set according to the road condition requiring a natural turn. desirable.

In this case, a steering wheel angle sensor, a yaw rate sensor or the like (not shown) is used as the turning detection means for detecting the turning state, and the window processing unit 24 (see FIG. 1) in the approaching object detection means 20 is used. ), The setting position of the window area may be changed in response to the turning signal from the turning detection means.

FIG. 6 is an explanatory diagram showing the window areas W1 to W7 when the vehicle turns right following a road having a right curve. The window areas W1 to W3 near the host vehicle turn right in the image G. Or set to be compressed. That is, the window processing unit 24 sets each of the window regions W1 to W7 on the assumption of the image G that relatively turns in accordance with the turning signal of the host vehicle.

Thus, by setting the window area according to the actual turning amount or turning speed by the steering wheel operation, the rear vehicle B can be optimally detected regardless of the road condition. Therefore, the approaching object detection means 20
Can accurately detect the movement of the rear vehicle B with respect to the own vehicle, and for example, can immediately generate the determination signal D when the rear vehicle B may collide with the own vehicle during turning. it can.

Example 7. Further, in order to easily identify the movement of the rear vehicle B, the wind regions W1 to Wn are set along the radiation road as shown in FIG. 2, but may be set along any direction according to the request. it can.

Although the case where the number n of the window regions W1 to Wn is 7 has been illustrated, it can be set to any number. At this time, the larger the number of window regions n is set, the more the determination accuracy of the rear vehicle B is improved. However, since the calculation time is correspondingly required, the calculation time of the representative value extraction unit 25, the shift amount calculation unit 28, etc. It is set to an appropriate value depending on the situation.

Example 8. Also, 1 for one display image
Set the window areas W1 to Wn of the block and set the typical values F1 to Fn.
Although only one data string of the pattern P is generated, the window regions W1 to Wn may be divided into a plurality of blocks and the pattern may be set for each block. In this case, the movement of the rear vehicle existing in each block can be individually divided and monitored with high accuracy.

Example 9. Further, the average value of the brightness of the pixels in the window area is adopted as the representative value F1 to Fn of each window area W1 to Wn, but the maximum value of the brightness, the minimum value, the difference between the maximum value and the minimum value, or , The ratio of the maximum value to the minimum value is the typical value F1 ~
It may be adopted as Fn. Especially at night, etc., because the headlights only light up when the headlights are turned on,
In order to make the feature amount of each window region W1 to Wn remarkable, it is preferable to set the maximum value to the representative value F1 to Fn rather than the average value of brightness.

Example 10. (Corresponding to claim 7) Further, the window regions W1 to Wn are set to one-dimensional to set each representative value F1.
Although the pattern P including the one-dimensional data string of Fn to Fn is created, the window region may be set to two-dimensional to create the pattern including the two-dimensional data string.

FIG. 7 shows a two-dimensional window region W11 to Wnm.
It is explanatory drawing which shows the image G by Example 6 which set (for example, W11-W76), and here seven pieces along the road direction,
An example is shown in which only six are set along the direction perpendicular to the road surface. FIG. 8 shows a two-dimensional pattern P by the rear vehicle B in FIG.
It is explanatory drawing which shows o'and P ', the pattern Po' of FIG.8 (a) is the last pattern at the time to, and the pattern P'of FIG.8 (b) is this time pattern at the time to + (DELTA) t.

FIG. 9 shows the shift correction amount ΔPf of the pattern P ′.
FIG. 9 is a characteristic diagram showing a relationship of a representative value deviation ΔF ′ with respect to n and ΔPfm, and a shift correction amount (ΔPfn, ΔPfm) that minimizes the representative value deviation ΔF ′ is obtained as a shift amount (ΔPn, ΔPm) composed of two-dimensional coordinates. To be

In this case, the shift amount calculating section 28 determines the pattern Po 'at a certain time to and the time to after the minute time Δt.
Shift amount with the pattern P'at + Δt (ΔPn, ΔPm)
Is calculated as shown in FIG. 9, the two-dimensional shift correction amount
The following representative value deviation ΔF ′ is repeatedly calculated while changing (ΔPfn, ΔPfm).

ΔF ′ = (1 / n) (1 / m) ΣΣ | f ′ (k, j) −F ′ (k + ΔPfn, j + ΔPfm) | ... (2)

However, in the equation (2), k = 1 to n, j =
1 to m, and f ′ (k, j) is the value of the previous pattern Po ′.
The representative value of the (k, j) th window region Wkj, F '(k +
ΔPfn, j + ΔPfm) is (k + ΔPfn,
j + ΔPfm) th representative value. Therefore, the representative value deviation Δ
F'is the shift correction amount (ΔPfn, ΔPf
It shows a value obtained by normalizing the total sum of the deviations of the representative values of the respective wind areas when they are shifted by m).

As shown in the equation (2), the characteristic diagram of FIG. 9 can be obtained by repeatedly calculating the shift correction amounts (ΔPfn, ΔPfm) in the radial direction and the vertical direction.
As a result, a two-dimensional shift amount (ΔPn, ΔPm) that minimizes the representative value deviation ΔF ′ is obtained, and this shift amount (ΔPn
n, ΔPm) is a value that best matches the previous pattern Po ′ and the current pattern P ′.

Therefore, the rear vehicle B has moved by the vector V in FIG. 9 during the minute time Δt, and similarly to the above, if the vector V indicates the positive direction, it is separated and the vector V is negative. If the direction is shown, it is determined that the vehicle is approaching.
As described above, by using the two-dimensional pattern P ′, the movement of the target object can be grasped in more detail, and the determination reliability of the approaching target object is improved.

[0079]

As described above, according to the first aspect of the present invention, the image pickup means for taking an image, the window processing section for dividing the image into a plurality of window regions, and the characteristic amount for each window region are representative values. A representative value extraction unit for extracting as, a pattern creation unit for creating a pattern corresponding to an image using each representative value as a data string, and a shift amount calculation for comparing each pattern of images at different imaging timings to calculate a shift amount Unit and a determination unit that generates a determination signal when the shift amount is equal to or greater than a predetermined value indicating the presence of the approaching object, the approaching object can be detected automatically and reliably. There is an effect that the device can be obtained.

According to a second aspect of the present invention, in the first aspect, the image pickup means is installed in the vehicle to capture a rear image of the vehicle, so that an approaching object from the rear of the vehicle is automatically detected. There is an effect that an approaching object detection device that can make a targeted and reliable determination can be obtained.

According to a third aspect of the present invention, in the second aspect, a controller for generating an alarm drive signal in response to the determination signal and an alarm means driven by the alarm drive signal are provided. Since the warning means is driven to warn the driver when an approaching object is determined, the approaching object that automatically and reliably determines the approaching object from the rear of the vehicle and ensures safety. An object detection device can be obtained.

According to a fourth aspect of the present invention, in the third aspect, there is provided switch means that is operated when the traveling direction of the vehicle is changed, and the imaging means displays a plurality of rear images corresponding to the traveling direction of the vehicle. An image is picked up, the determination unit generates a determination signal for each rear image, and the controller generates an alarm drive signal when the determination signal for the changed traveling direction is detected in response to the operation signal from the switch means. As a result, the approaching object detecting device that automatically and surely determines the approaching object on the changed traveling path and notifies the danger to ensure the safety can be obtained.

According to claim 5 of the present invention, in any one of claims 2 to 4, there is provided turning detection means for detecting a turning state of the vehicle, and the window processing part is provided with the turning detection means. In response to the turning signal, the set position of the window area is changed to detect the backward object that is approaching on the road after turning, so the approaching object is automatically and reliably determined regardless of the road condition. Therefore, there is an effect that an approaching object detection device that secures safety can be obtained.

According to claim 6 of the present invention, in any one of claims 1 to 5, the window region is set along the radiation from the gazing point at infinity, and the representative value pattern is one-dimensional. Since the data sequence is set to (1), there is an effect that the approaching object detection device that can automatically and surely determine the approaching object is obtained.

According to claim 7 of the present invention, in any one of claims 1 to 5, the window region is set along the radiation from the gazing point at infinity, and Since the representative value pattern is set along the vertical direction and the representative value pattern is a two-dimensional data string, there is an effect that the approaching object detection device that can determine the approaching object automatically and surely can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing a functional configuration of an approaching object detection means according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a window area set according to the first embodiment of the present invention together with an image.

FIG. 3 is an explanatory diagram showing a pattern created according to the first embodiment of the present invention.

FIG. 4 is a characteristic diagram showing changes in the representative value deviation with respect to the shift correction amount calculated according to the first embodiment of the present invention.

FIG. 5 is a block diagram showing a functional configuration of a fifth embodiment of the present invention.

FIG. 6 is an explanatory diagram showing a window area set according to a sixth embodiment of the present invention.

FIG. 7 is an explanatory diagram showing a window area set according to Example 10 of the present invention together with an image.

FIG. 8 is an explanatory diagram showing a pattern created according to Example 10 of the present invention.

FIG. 9 is a characteristic diagram showing changes in the representative value deviation with respect to the shift correction amount calculated according to the tenth embodiment of the present invention.

FIG. 10 is a block diagram showing an example of a conventional approaching object detection device.

11 is an explanatory diagram showing an approaching object detection operation by the apparatus of FIG.

FIG. 12 is an explanatory diagram showing an operation of the second example of the conventional approaching object detection device.

FIG. 13 is an explanatory diagram showing a display image by a third example of the conventional approaching object detection device.

[Explanation of symbols]

 21 camera (imaging means) 24 window processing section 25 representative value extraction section 26 pattern creation section 28 shift amount calculation section 29 determination section 31 winker switch (switch means) 32 controller 33 alarm means G image B rear vehicle D determination signal E alarm drive Signal S Operation signal F1 to Fn Representative value P, P'Current pattern Po, Po 'Previous pattern ΔP, ΔPn, ΔPm Shift amount W1 to Wn Wind area

Claims (7)

[Claims] 1. An image pickup unit for photographing an image, a window processing unit for dividing the image into a plurality of window regions, a representative value extraction unit for extracting a feature amount of each window region as a representative value, and each of the representatives. A pattern creating unit that creates a pattern corresponding to the image using a value as a data string, a shift amount calculating unit that compares each pattern of the image at different imaging timings to calculate a shift amount, and the shift amount is a predetermined value. An approaching object detection device, comprising: a determination unit that generates a determination signal in the above case, wherein the predetermined value is set to a level indicating the presence of an object approaching the imaging unit. 2. The approaching object detection device according to claim 1, wherein the imaging means is installed in a vehicle and captures a rear image of the vehicle. 3. The approaching object detection device according to claim 2, further comprising a controller that generates an alarm drive signal in response to the determination signal, and an alarm unit that is driven by the alarm drive signal. 4. A switch unit that is operated when the traveling direction of the vehicle is changed, wherein the imaging unit captures a plurality of rear images corresponding to the traveling direction of the vehicle, and the determination unit includes the rear portions. Generating a determination signal for the image, wherein the controller generates the alarm drive signal when a determination signal for the changed traveling direction is detected in response to an operation signal from the switch means. The approaching object detection device according to claim 3. 5. A turning detection unit for detecting a turning state of the vehicle is provided, and the window processing unit changes a set position of the window region in response to a turning signal from the turning detection unit. The approaching object detection device according to any one of claims 2 to 4. 6. The window region is set along a ray from a gazing point at infinity, and the pattern comprises a one-dimensional data string. Approaching object detection device. 7. The window area is set along a ray from a gazing point at infinity and is set along a direction perpendicular to the ray, and the pattern is formed from a two-dimensional data string. The approaching object detection device according to any one of claims 1 to 5, wherein: