JP3738949B2 - Vehicle peripheral area monitoring device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a vehicle peripheral area monitoring device, and in particular, a vehicle approaching from the rear and rear sides using an image obtained by imaging the rear (including the rear side) by a CCD camera (solid-state imaging device camera) installed in the vehicle, The present invention relates to a vehicle surrounding area monitoring device that detects an obstacle and gives a warning to a driver.
[0002]
[Prior art]
When changing lanes to adjacent lanes while driving a vehicle, if you overlook a vehicle that is traveling at a higher speed than the host vehicle or an obstacle (such as a stopped vehicle or construction work) in the adjacent vehicle, There is a risk of accidents occurring in either case, such as when sudden braking is applied without knowing that the following vehicle in the same lane has suddenly approached.
For this reason, a vehicle peripheral area monitor has been conventionally used to perform image processing on images captured rearward and rearward using a CCD camera to recognize the approaching vehicle and obstacles and to give a warning to the driver. Various devices have been proposed.
[0003]
For example, in Japanese Patent Application Laid-Open No. 7-50769 (Prior Art 1), when a rear image of a CCD camera (video camera) is captured from the rear part of a traveling vehicle, the infinite distance increases as the rear approaching object moves away by forward traveling of the vehicle. Focusing on the point moving in the direction of focusing on the point, extracting one point of interest of the image captured at every Δt time interval, and obtaining the pixel movement amount connecting the points of the previous image and the subsequent image as an optical flow Then, the speed vector component of the optical flow is calculated, and the approaching vehicle is recognized in comparison with the still image.
That is, a still image such as a scene or a mark on the road surface has a vector length corresponding to the traveling speed of the own vehicle and heads to an infinite point, but in the case of an approaching vehicle, the vector length depends on the relative speed with respect to the own vehicle. However, it is possible to determine the approaching state of the vehicle or the like by comparing the vector lengths.
[0004]
In Japanese Patent No. 2701651 (Prior Art 2), the left and right target areas corresponding to the traveling lane are searched within the vehicle monitoring band in the rear image captured by the CCD camera (TV camera), and the respective areas are searched. A vehicle image located in (travel lane) is set as a presence candidate, and by detecting a state in which the vehicle image moves left and right and in the center between the lanes, it is possible to determine the approaching state of the vehicle and the risk of a collision.
[0005]
[Problems to be solved by the invention]
In any of the above techniques, in order to improve the monitoring accuracy and avoid the risk of collision, the CCD camera has a wide angle of view to monitor a wide area from the rear left side to the center to the right side, in other words, the CCD pixel. Must be scanned in a wide range to perform image processing.
However, when the monitoring area of the CCD image becomes large, image processing such as density binarization conversion and calculation processing for each pixel becomes complicated, and processing time is required. As a result, prediction of collision or the like cannot be performed in a short time, There was a big problem in actual operation.
In order to eliminate such a drawback, it is possible to reduce the processing time by performing a thinning process in the monitoring area, but this deteriorates the detection performance and the detection accuracy.
[0006]
In view of the shortcomings of the prior art, the present invention collects the trajectory information of the detection target vehicle in advance and sets a region with a high detection frequency as a monitoring region, thereby reducing the monitoring region and improving detection accuracy such as thinning processing. A vehicle that can reduce the processing time while maintaining the detection accuracy necessary for detecting a collision, etc. without performing a lowering operation, and thereby can determine the approaching state of the vehicle and the risk of a collision accurately and quickly. An object is to provide a peripheral area monitoring device.
[0007]
[Means for Solving the Problems]
In order to solve this problem, the present invention provides an invention according to claim 1,
An imaging unit provided in a vehicle for imaging a predetermined direction from the vehicle, a monitoring region setting unit for setting a monitoring region from an image obtained from the imaging unit, and performing image processing by thinning out the set monitoring region The first image processing means to perform, the trajectory data collection means for collecting the trajectory data of the monitored object from the first image processing means, and the trajectory data exists based on the trajectory data obtained from the trajectory data collection means. whether the a risk prediction area setting means for setting a monitoring area as a risk prediction region, and a second image processing means for performing a normal or dense image processing in the danger prediction area, said whether the vehicle is in a condition to be alert to an alarm condition determining means for determining whether the second image processing means in place of the first image processing means and the conditions to be a warning is established based on a determination from said alarm condition judging means To provide a vehicle surrounding area monitoring device characterized by comprising a selection means for-option.
[0008]
Further, the invention according to claim 2 relates to claim 1, and when the second image processing means is selected by the selection means, an alarm determination is made based on image processing data from the second image processing means. If the alarm is determined , the alarm means is activated .
[0009]
Here, a signal (alarm condition signal) that is a basis of a condition to be alarmed includes a direction indicator (blinker) signal, a steering angle signal, a brake signal, an accelerator signal, a turbo signal, a speed sensor, a fog lamp signal, and the like.
Also, the danger prediction area is the rear side corresponding to the adjacent lane to bend in the case of lane change based on the direction indicator signal or steering angle signal, the rear front in the case of sudden braking based on the brake signal, and the accelerator signal, In the case of rapid acceleration based on a turbo signal and a speed sensor, it is ahead. Therefore, the monitoring area indicates not only the rear but also one or a plurality of areas of the rear side, the left and right sides, the front, and the front side.
[0010]
Incidentally, a technology that uses a turn signal signal as a detection signal of a vehicle collision prevention device is disclosed in Japanese Patent No. 2641562. However, this technology simply takes an AND with the following vehicle detection means to check whether there is an alarm. Is a technique different from the present invention.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, the present invention will be described in detail with reference to the embodiments shown in the drawings. However, unless otherwise specified, the dimensions, shapes, relative arrangements, and the like of the components described in this embodiment are merely illustrative examples and not intended to limit the scope of the present invention.
FIG. 1 is a block diagram of a vehicle peripheral area monitoring apparatus according to an embodiment of the present invention, and FIG. 2 is a flowchart thereof. In the present embodiment, the direction indicator signal is used as an alarm condition signal, and the monitoring device is intended to prevent a collision with the following vehicle in the case of a lane change.
1 and 2, reference numeral 1 denotes a CCD camera (charge coupled device camera) that widens the angle of view and captures images from the inside of the vehicle to the rear side of the vehicle and the rear side of the adjacent lane. This corresponds to the image input unit in (S1).
[0012]
Since the rear image picked up by the CCD camera 1 is analog data, the digital gradation image is converted into a digital gradation image by the AD converter 2, and then the gray level conversion means in the binarization conversion circuit 3 is converted into the digital gradation image. After extracting a black level having a predetermined length or less, the floating binarization means converts the black level into binary image data, which is then developed in the video memory 4.
[0013]
In the alarm condition determination unit (alarm condition determination means) 5, when the alarm condition is not satisfied (S2) (for example, when the direction indicator is turned on and the alarm condition is satisfied , the direction indicator is turned off. By transmitting a wide area selection signal to the circuit 6 and masking unnecessary portions of the image developed in the video memory 5, for example, a wide monitoring area 60 that can be detected by the following vehicle in the adjacent lane area. Set (S3).
In this case, the masking can be arbitrarily set, but in the case of the present embodiment, only the portion corresponding to the own vehicle traveling lane on the rear rear side of the own vehicle is masked because it is the rear monitoring in the case of lane change. The left and right adjacent lane areas and the portions corresponding to the guard rails on the outside are entered into the monitoring area.
[0014]
Further, even on roads where the driving lane is not clear or roads where there is no white line, such as snowy roads, the masking range may be limited and a wide monitoring area may be set including the phage portion.
In this way, by setting the monitoring area as a wide area, it is possible to capture not only the road and snow road where the travel lane is not clear, but also the entire adjacent lane area even when the imaging road is a curved curve.
[0015]
Next, while thinning out the monitoring area by the thinning processing circuit 7 (S4), the arithmetic circuit 8 calculates and collects optical flow data (S5). An example of the calculation and collection of the optical flow data is also described in the prior art 1, and a detailed description thereof will be omitted.
Based on the flow data, the trajectory of the monitoring object such as the following vehicle or the obstacle is recorded as data in the trajectory memory 9 (S6).
[0016]
In the case of the first routine, the monitoring area is thinned out. However, since this routine is not instantaneous determination data but data collection for long-term trajectory collection, pixel data after thinning processing is performed. Is enough.
[0017]
On the other hand, when the warning condition is satisfied (for example, turn signal indicator) as in the case of changing lanes (S11), the process proceeds to the second routine, and is recorded when the warning condition is not satisfied. The trajectory data of the monitored object such as the following vehicle is taken into the aperture region setting circuit 11, and the setting circuit 11 performs image recognition of the trajectory data (approaching vehicle) in comparison with the still image, and performs the masking process to detect the trajectory data. The monitoring area is narrowed down to the area where the error exists (S12).
As a result, the monitoring area can be narrowed down to an area where the most objects to be alarmed appear, that is, an area where an accident such as a collision is likely to occur due to an approaching vehicle or the like.
[0018]
Then, the image accumulation data of the narrowed area is taken into the alarm determination unit 12, and the alarm determination unit 12 performs the alarm determination by normal image processing without performing the thinning process as in the first routine (S13). ).
If the determination unit 12 determines that an alarm signal is present, that is, an alarm display is required , the alarm display output unit 13 outputs a predetermined alarm (S14).
[0019]
【The invention's effect】
As described above, according to the present invention, it is possible to reduce the processing time for monitoring the extra area by concentrating and monitoring the area where the probability that the object to be monitored appears is high. Since the inside is not thinned out, the accuracy detection performance can be improved.
Also, as shown in the preceding example 2, compared to a monitoring area setting method in which a driving lane is recognized and a monitoring area is set, it is effective even on roads where the driving lane is not clear or roads without white lines such as snowy roads. The monitoring area can be set.
[Brief description of the drawings]
FIG. 1 is a block diagram of a vehicle surrounding area monitoring apparatus according to an embodiment of the present invention.
FIG. 2 is a flowchart showing the operation procedure of FIG. 1;
[Explanation of symbols]
1 CCD camera (charge-coupled device camera)
3 Binarization conversion circuit 4 Video memory 5 Alarm condition determination unit 6 Area selection circuit 7 Thinning-out processing circuit 9 Trajectory memory 11 Aperture area setting circuit 12 Alarm determination unit 13 Alarm display output unit

Claims (2)

An imaging means provided in the vehicle and imaging a predetermined direction from the vehicle;
Monitoring area setting means for setting a monitoring area from an image obtained from the imaging means;
First image processing means for performing image processing by thinning out the set monitoring area;
Trajectory data collection means for collecting trajectory data of the monitored object from the first image processing means;
Danger prediction area setting means for setting the monitoring area where the locus data exists based on the locus data obtained from the locus data collection means as a danger prediction area;
A second image processing means for performing normal or precise image processing in the danger prediction area ;
Alarm condition determining means for determining whether or not the vehicle is in a condition to be alarmed;
A vehicle comprising: selection means for selecting the second image processing means instead of the first image processing means when a condition to be warned is established based on the determination from the warning condition determination means Perimeter monitoring device. When the second image processing means is selected by the selection means, an alarm is determined based on the image processing data from the second image processing means, and when the alarm is determined, the alarm means is activated. The vehicle surrounding area monitoring apparatus according to claim 1, wherein

Images