JP2001294109A - Rear monitoring device for vehicle and its controlling method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suitably monitor a vehicle or the like approaching one's own vehicle. SOLUTION: A differential value of the concentration of each pixel of an omnidirectional image by an omnidirectional visual sensor is calculated by a projection part; a concentration differential image is led out and an annular watching region is set in it; an annular edge histogram is formed by calculating a concentration projection value that is an added sum of concentration in the radiation direction in the watching region by the projection part; a time series image expressing the temporal variation of the concentration projection value is formed by developing the edge histogram into a band-like form by a binarization part with the six-o'clock direction as a reference position; a binarization image is formed by binarizing the time series image by a predetermined reference concentration threshold value; whether a region having higher concentration than the reference concentration threshold value whose width continuously and gradually increases in terms of time is present or not in the binarization image is determined by a determination part; and when it is present, the high concentration region is recognized as an obstacle approaching the one's own vehicle from the rear.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicular rear monitoring device for processing an image behind a vehicle and monitoring a vehicle approaching the vehicle, and a control method therefor.

[0002]

2. Description of the Related Art Conventionally, back sonars and corner sonars as shown in FIGS. 9 and 10 have been known as devices for monitoring the rear of a vehicle. The back sonar emits ultrasonic waves rearward by an ultrasonic sensor provided at the rear part of the vehicle B, and detects the presence of an obstacle behind by detecting reflected waves from the obstacle. On the other hand, the corner sonar emits ultrasonic waves obliquely forward and obliquely backward by ultrasonic sensors provided at the corners of the vehicle B, and detects reflected waves from obstacles, thereby obstructing obliquely forward and obliquely backward. It detects the presence of an object.

[0003]

However, in these conventional devices, since the direction and area to be monitored are all limited, a large number of sensors must be provided for monitoring at a wide angle. There is a disadvantage that the configuration is complicated and expensive.

The sensitivity of the ultrasonic sensor is at most 4 to
Since it is about 5 m, it is particularly suitable for rearward monitoring during parking, but is not suitable for detecting vehicles approaching from behind or behind the vehicle during traveling, and is used as an aid for safe driving. Is not enough.

Accordingly, an object of the present invention is to enable a vehicle or the like approaching a vehicle to be accurately monitored.

[0006]

In order to achieve the above-mentioned object, a vehicular rear monitoring apparatus according to the present invention comprises: a convex mirror for converging light reflected from an object to project a reflected image of the surroundings; An omnidirectional vision sensor mounted on a vehicle, comprising imaging means for capturing an image reflected by the convex mirror and outputting an image signal, and an image input section for forming an omnidirectional image by receiving an image signal from the omnidirectional vision sensor. And derives a density differential image of the omnidirectional image, sets an annular gaze area for the density differential image, calculates a radial density projection value in the gaze area in the radial direction, and forms an annular edge histogram. A projection unit for forming a time-series image representing a temporal change of a density projection value in a state where the annular edge histogram is developed from a reference position, and forming a binary image with a predetermined reference density threshold A high-density area when the width of the high-density area from the binarized image obtained by the binarization section becomes larger gradually and continuously over time than the reference density threshold. And a judging unit for judging that the obstacle approaches the own vehicle from behind.

According to such a configuration, the annular edge histogram derived from the density differential image of the omnidirectional image is developed from the reference position, and the time-series image representing the temporal change of the density projection value of the developed edge histogram is obtained. Is binarized by a predetermined reference density threshold, and when the high density area having the predetermined density threshold or more continuously increases in width over time, the high density area is automatically replaced from the rear. It is determined that the obstacle is approaching the car.

[0008] Therefore, wide-angle monitoring can be performed without arranging a large number of sensors as in the related art, and in particular, the presence or absence of an obstacle approaching from behind can be accurately determined.

Further, in the vehicle rear monitoring device according to the present invention, when the width of the high-density region higher than the reference density threshold corresponds to the width of the vehicle, the determining unit may detect the obstacle from behind. It is characterized by determining that the vehicle is approaching.

According to this configuration, when the width of the high-density region corresponds to the width of the vehicle, the approaching obstacle is determined to be the vehicle, so that the vehicle approaching the vehicle from behind can be reliably determined. Can be recognized.

Further, the vehicle rear monitoring apparatus according to the present invention is characterized in that it comprises an alarm section for issuing an alarm when the existence of an obstacle approaching the own vehicle is judged by the judgment section.

[0012] According to this configuration, the driver can easily recognize that there is an obstacle such as a vehicle approaching the vehicle from behind based on the warning of the warning unit, and quickly take necessary measures. Can be adopted.

Further, the control method of the vehicular rear monitoring apparatus according to the present invention is characterized in that an omnidirectional vision sensor is mounted on a vehicle to form an omnidirectional image around the vehicle, and a density differential image of the omnidirectional image is formed. An annular gaze area is set for the derived density differential image, a density projection value in the radial direction in the gaze area is calculated, an annular edge histogram is formed, and this is developed from the reference position and developed. A time-series image representing a temporal change of the density projection value is formed in the edge histogram, and is binarized by a predetermined reference density threshold to form a binarized image. When the width of the high-density region becomes larger than the threshold value over time, the region is determined to be an obstacle approaching the vehicle from behind.

According to such a configuration, wide-angle monitoring can be performed without arranging a large number of sensors, and particularly, the presence or absence of an obstacle approaching from behind can be accurately determined.

Further, in the control method of the vehicle rear monitoring device according to the present invention, when the width of the high density area higher than the reference density threshold value corresponds to the width of the vehicle, the obstacle is approached from behind. It is characterized by being determined to be a vehicle. According to such a configuration, a vehicle approaching the vehicle from behind can be reliably recognized.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described with reference to FIGS. However, FIG. 1 is a block diagram, FIG. 2 is a cross-sectional view showing a schematic configuration of a part, and FIGS.

In FIG. 1 showing the overall configuration, reference numeral 1 denotes an omnidirectional visual sensor mounted on a vehicle mounted on the center of the ceiling of the vehicle interior or near the upper part of the rear window, and 2 denotes an image signal from the omnidirectional visual sensor 1. An image input unit 3 which is input and forms an omnidirectional image (see FIG. 4) is a projection unit, which derives a density differential image of the omnidirectional image by the image input unit 2 and performs an annular operation on the density differential image. (See R in FIG. 5) is set, and the density projection value in the radial direction in the focused area is calculated to form a circular edge histogram (see FIG. 6).

Further, in FIG. 1, reference numeral 4 denotes a binarizing unit,
The annular edge histogram is developed in a band shape using the 6 o'clock direction of the annular edge histogram (the forward direction in the omnidirectional image in FIG. 4) as a reference position, and a time-series image representing the temporal change of the density projection value is formed. The sequence image is binarized with a predetermined reference density threshold to form a binarized image.

In FIG. 1, reference numeral 5 denotes a judgment unit;
In the binarized image obtained by the binarizing unit 4, when there is a high-density region that is larger than a reference density threshold whose width continuously increases gradually over time, the high-density region is positioned from the rear. It is determined that the obstacle is approaching the vehicle. At this time, when the width of the high-density area higher than the reference density threshold corresponds to the width of the vehicle, the determination unit 5 determines that the obstacle is a vehicle approaching from behind.

In FIG. 1, reference numeral 6 denotes an alarm unit. When the judgment unit 5 judges that there is an obstacle approaching the vehicle from behind, the driver approaches the vehicle from behind. Gives an alarm that an obstacle is present. The alarm at this time is desirably, for example, lighting or blinking of a warning lamp provided on the instrument panel.

As shown in FIG. 2, for example, the omnidirectional vision sensor 1 includes a case 1a, an imaging means 1b housed in the case 1a, and a cylindrical transparent body 1c connected to the upper part of the case 1a. , A lid 1d closing the upper opening of the transparent body 1c, and a convex mirror 1e attached inside the lid 1d.

At this time, the convex mirror 1e is composed of a spherical mirror, a paraboloid of revolution, a hyperboloid of revolution, or the like, and condenses the reflected light from the surrounding object to project a reflected image of the surrounding. The imaging means 1b includes a two-dimensional image sensor such as a CCD camera and a signal from the two-dimensional image sensor.
It is composed of an image processing circuit that performs conversion processing to a composite video signal conforming to standards such as AL. Then, the image reflected by the convex mirror 1e is picked up by the image pickup means 1b, and the composite video signal is output to the subsequent image input section 2 as an image signal. Note that the one-dot chain line arrow in FIG. 2 indicates the line of sight of the imaging unit 1b.

By the way, the principle of the cylindrical projection conversion in the image input unit 2 will be briefly described. As shown in FIG. 3, a virtual cylindrical surface having a radius Rp is set around the Z axis, and each of the omnidirectional images is set. This is possible by projecting the mapping points on a cylindrical surface. Now, FIG. 3 is a view of the hyperboloid mirror viewed from the horizontal direction, and a mapping point of a point P (X, Y, Z) in a three-dimensional space on an omnidirectional image is defined as p (x, y). Assuming that the distance between p and the center of the image is r, the azimuth of the point P is θ, and
The following relationship is established.

[0024]

(Equation 1)

On the other hand, the hyperboloid is a two-lobed hyperboloid obtained by rotating the hyperbola about the Z axis.
When, hyperboloidal mirror has two focal (0,0, c) and (0,0, -c) ^{(where, c 2 = a 2 + b} 2).
Here, a, b, and c are parameters of the hyperboloid.

[0026]

(Equation 2)

Then, the center of the imaging lens of the imaging means 1b (see FIG. 2) is set to a focal position on the Z <0 side, and an omnidirectional image with respect to an arbitrary point P (X, Y, Z) in a three-dimensional space is set. Assuming that the above-mentioned mapping point is p (x, y), the relationship of Expression 3 holds between the point P and the mapping point p. Note that in Equation 3, f is the focal length of the camera lens.

[0028]

(Equation 3)

Therefore, from the above equations 1 and 3,
By determining the height Z and the direction angle θ of the point on the cylindrical surface,
The corresponding image coordinates (x, y) can be uniquely determined, and the omnidirectional image can be converted into a cylindrical projection image.

The cylindrical projection image obtained in this way is
Since there is almost no distortion as in the case of a normal captured image, it is very easy to see when displayed on a monitor, and the distance and positional relationship between the vehicle and its surroundings can be clearly and easily understood.

Now, when an omnidirectional image as shown in FIG. 4 is obtained by the image input unit 2 based on the image signal from the omnidirectional visual sensor 1, the projecting unit 3 performs a process for each pixel of the omnidirectional image. The derivative of that concentration is calculated,
A density differential image as shown in FIG. 5 is derived, and an annular fixation region R is generated for the derived density differential image.
(See FIG. 5) is set.

Further, the projection unit 3 calculates a density projection value, which is the sum total of the densities in the radial direction in the circular attention area R set in the density differential image, and obtains a circular edge as shown in FIG. A histogram is formed. FIG.
, Each peak point indicates that the density projection value in the radial direction is high.

Subsequently, the binarizing unit 4 develops the annular edge histogram shown in FIG. 6 into a band shape with the 6 o'clock direction as a reference position (arrow in FIG. 6), and obtains the density projection values as shown in FIG. Is formed, and the time-series image is binarized with a predetermined reference density threshold to form a binarized image as shown in FIG.

Then, in such a binarized image (see FIG. 8), if there is a higher density area than the reference density threshold whose width gradually increases continuously over time,
The determination unit 5 determines that the high-density area is an obstacle approaching the vehicle from behind, and when the width of the high-density area corresponds to the width of the vehicle, the obstacle approaches from behind. The vehicle is determined to be a vehicle, and a warning is issued by the warning unit 6.

Here, for example, when the own vehicle is moving backward to enter the garage or change direction, the stationary object at the rear will relatively approach the own vehicle direction. Can also detect a stationary object in the rear as an obstacle approaching the own vehicle, and the alarm unit 6 issues an alarm to the driver.

Therefore, according to the above-described embodiment, it is not necessary to dispose a large number of sensors as in the related art, and it is possible to process images from a single omnidirectional vision sensor 1 and perform wide-angle monitoring. In addition, even while the vehicle is running, it is possible to reliably determine whether there is an approaching vehicle or another obstacle from behind.

When a vehicle approaching from behind is present, a warning is issued by the warning unit 6 to notify the driver of the warning. It is possible to easily recognize that there is and take necessary measures.

In the above-described embodiment, a case has been described in which the 6 o'clock direction of the omnidirectional image (the forward direction in the omnidirectional image) is used as the reference position when developing the annular edge histogram. Of course, it is not limited to this.

Furthermore, in the above-described embodiment, the case has been described where all the alarms are issued in the same manner by the alarm unit 6 when the judgment unit 5 judges that there is an obstacle approaching the vehicle from behind. , Change the mode of alarm between when the obstacle is a vehicle and when it is not,
In particular, in the case of a vehicle, the driver may be alerted.

In the above-described embodiment, the case where the warning by the warning unit 6 is the lighting or blinking of the warning lamp has been described. It goes without saying that a buzzer may be sounded, or a predetermined warning message may be output as voice by voice synthesis means.

The present invention is not limited to the above-described embodiment, and various changes other than those described above can be made without departing from the gist of the present invention.

[0042]

As described above, according to the first and fourth aspects of the present invention, the annular edge histogram derived from the density differential image of the omnidirectional image is developed from the reference position, and A time-series image representing a temporal change in the density projection value is binarized by a predetermined reference density threshold, and a high-density area having a predetermined density threshold or higher is successively temporally increased in width. When it comes to the high concentration area, it is determined as an obstacle approaching the vehicle from behind, so that wide-angle monitoring can be performed without disposing many sensors as in the past, and especially approaching from behind. This makes it possible to accurately determine the presence or absence of an obstacle, and is very effective not only for parking but also for monitoring the rear of the vehicle while the vehicle is traveling, and is suitable as a vehicle driving assistance.

According to the second and fifth aspects of the present invention, when the width of the high-density region corresponds to the width of the vehicle, the approaching obstacle is determined to be a vehicle. The vehicle approaching the vehicle can be reliably recognized.

According to the third aspect of the present invention, the driver can easily recognize that there is a vehicle approaching the vehicle from behind based on the warning of the warning unit, and can quickly perform the necessary operations. Actions can be taken.

[Brief description of the drawings]

FIG. 1 is a block diagram of one embodiment of the present invention.

FIG. 2 is a sectional view showing a schematic configuration of an embodiment of the present invention.

FIG. 3 is an operation explanatory diagram of one embodiment of the present invention.

FIG. 4 is an operation explanatory diagram of one embodiment of the present invention.

FIG. 5 is an operation explanatory diagram of one embodiment of the present invention.

FIG. 6 is an operation explanatory diagram of one embodiment of the present invention.

FIG. 7 is an operation explanatory diagram of one embodiment of the present invention.

FIG. 8 is an operation explanatory diagram of one embodiment of the present invention.

FIG. 9 is a schematic view of a conventional example.

FIG. 10 is a schematic view of another conventional example.

[Explanation of symbols]

 Reference Signs List 1 omnidirectional visual sensor 1b imaging means 1e convex mirror 2 image input unit 3 projection unit 4 binarization unit 5 judgment unit 6 alarm unit

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06T 1/00 330 G06T 1/00 330B G08G 1/16 G08G 1/16 C H04N 7/18 H04N 7/18 J

Claims (5)

[Claims] 1. A vehicular rear monitoring device that processes an image behind a host vehicle and monitors a vehicle approaching the host vehicle, and collects reflected light from an object to be imaged to project a surrounding reflected image. An omnidirectional vision sensor mounted on the vehicle, comprising a convex mirror and an image pickup means for capturing an image reflected by the convex mirror and outputting an image signal; and forming an omnidirectional image by inputting the image signal from the omnidirectional vision sensor. An image input unit, which derives a density differential image of the omnidirectional image, sets an annular gazing area for the density differential image, calculates a radial density projection value in the gazing area in a radial direction, and obtains an annular edge histogram. Forming a time-series image representing a temporal change of a density projection value in a state where the annular edge histogram is developed from a reference position, and forming a predetermined reference density threshold A binarizing unit for performing binarization, and when the width of the high-density region from the binarized image obtained by the binarizing unit becomes gradually larger in time than the reference density threshold, the binary A determination unit configured to determine a high concentration region as an obstacle approaching the vehicle from behind. 2. The method according to claim 1, wherein the determining unit determines that the obstacle is a vehicle approaching from behind when the width of the high-density area higher than the reference density threshold corresponds to the width of the vehicle. The vehicle rear monitoring device according to claim 1. 3. The rear monitor for a vehicle according to claim 1, further comprising an alarm unit that issues an alarm when the presence of an obstacle approaching the own vehicle is determined by the determination unit. apparatus. 4. A control method of a vehicle rear monitoring apparatus for processing an image behind a host vehicle and monitoring a vehicle approaching the host vehicle, comprising: mounting an omnidirectional visual sensor on the vehicle; Forming an image, deriving a density differential image of the formed omnidirectional image, setting an annular gazing area for the density differential image, calculating a density projection value in the radial direction in the gazing area, and forming an annular An edge histogram is formed and developed from a reference position, a time-series image representing a temporal change of a density projection value is formed in the developed edge histogram, and the time series image is binarized by a predetermined reference density threshold value. An object which forms a binarized image and which approaches the host vehicle from behind when the width of the high-density area becomes larger than the reference density threshold from the binarized image continuously and gradually over time. Judge The method of the back monitoring apparatus for a vehicle according to claim and. 5. The vehicle according to claim 4, wherein when the width of the high-density area higher than the reference density threshold value corresponds to the width of the vehicle, the obstacle is determined to be a vehicle approaching from behind. Control method of the vehicle rear monitoring device.