JP2016110484A - Approaching vehicle detection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To surely specify a lane in which an approaching vehicle is traveling.SOLUTION: When a head lamp detection part 70 detects a head lamp of other vehicle, from an original image I(t) including a rear side road surface of a vehicle 10 imaged by a rear camera 50 (imaging part), a luminance distribution calculation part 73a calculates average luminances B(t), B(t) of inside of detection ranges W1, W2 (windows) of predetermined size, set by a window setting part 72 in an overview image in which the original image I(t) is overview converted, and an other vehicle travel lane determination part 64 determines whether, other vehicle travels on a lane adjacent to a lane in which the vehicle 10 travels, or, on a lane adjacent to the lane which is adjacent to the lane in which the vehicle travels, based on the average luminances B(t)(B(t)).SELECTED DRAWING: Figure 3

Description

  The present invention relates to an approaching vehicle detection device that detects an approaching vehicle that is installed in a vehicle and is traveling in an adjacent lane when the lane is changed, and issues an alarm.

  In recent years, a rear side monitoring function (BSW (Blind Spot) that installs a camera on a vehicle and detects an approaching vehicle from an image obtained by imaging the rear side of the vehicle that becomes a blind spot with respect to the driver and alerts the vehicle. Warning)) technology has been proposed.

  For example, in the rear side vehicle alarm device described in Patent Document 1, a moving object present in an adjacent lane adjacent to a traveling lane in which the vehicle is traveling, and a lane marker far from the vehicle in the adjacent lane And a method of detecting another vehicle traveling in an adjacent lane based on these detection results has been proposed.

Japanese Patent Laid-Open No. 9-240397

  However, in such a conventional rear side vehicle alarm device, for example, at night, if the light of the headlamps of other vehicles traveling in two adjacent lanes is reflected on the road surface, An image is captured as if another vehicle is present in the lane. For this reason, there is a risk of erroneous detection that there is another vehicle traveling in the adjacent lane.

  The present invention has been made in view of the above problems, and other vehicles that are traveling in two adjacent lanes of a traveling lane in which the vehicle is traveling are traveling in adjacent lanes adjacent to the traveling lane. An object of the present invention is to provide an approaching vehicle detection device that is not erroneously detected as a vehicle.

  In order to solve the above-mentioned problem, an approaching vehicle detection device according to the present invention is a vehicle that is approaching from the rear by turning on a headlamp in an adjacent lane of a lane in which the vehicle is traveling from a traveling vehicle. An approaching vehicle detection device for detecting and notifying an image, wherein an imaging unit that captures an image including a road surface on the rear side of the vehicle, and a position corresponding to a rear side region of the vehicle in the image A window setting unit for setting at least one window of a predetermined size having a predetermined positional relationship with the vehicle, and determining whether a headlamp of another vehicle is detected in a predetermined area of the image When the headlamp detection unit and the headlamp detection unit detect a headlamp of another vehicle, the luminance evaluation in each window is performed based on the luminance value of the area corresponding to each window of the image. A luminance distribution calculation unit for calculating values, and Based on the brightness evaluation value in the window, the other vehicle is present in an adjacent lane adjacent to the lane in which the vehicle is traveling, or is adjacent to the adjacent lane on the far side of the vehicle. And an other vehicle travel lane determination unit that determines whether the vehicle exists in an adjacent lane.

  According to the approaching vehicle detection device according to the present invention configured as described above, it is possible to reliably determine whether the approaching vehicle is traveling in the adjacent lane or the adjacent adjacent lane by adopting the above configuration. can do. Therefore, since the approaching vehicle traveling in the adjacent lane is not erroneously detected as the approaching vehicle traveling in the adjacent lane, the presence of the approaching vehicle can be more accurately notified.

It is a 1st figure explaining the road environment to which the approaching vehicle detection apparatus which is one Embodiment of this invention is applied. It is a 2nd figure explaining the road environment to which the approaching vehicle detection apparatus which is one Embodiment of this invention is applied. It is a block diagram which shows the hardware constitutions in the approaching vehicle detection apparatus of Example 1 which is 1 embodiment of this invention. It is a functional block diagram which shows the function structure in the approaching vehicle detection apparatus of Example 1 which is 1 embodiment of this invention. In Example 1, it is a figure explaining the process which detects the headlamp of another vehicle. In Example 1, it is a figure explaining the example of the window which a window setting part sets. In Example 1, it is the 1st figure explaining the method of determining the traveling lane of another vehicle by the luminance distribution in a window. In Example 1, it is the 2nd figure explaining the method of determining the traveling lane of another vehicle by the luminance distribution in a window. 3 is a flowchart illustrating a flow of a series of processes performed in the first embodiment. It is a flowchart which shows the detailed flow of the other vehicle travel lane determination process shown in the flowchart of FIG. It is a functional block diagram which shows the function structure in the approaching vehicle detection apparatus of Example 2 which is 1 embodiment of this invention. In Example 2, it is a figure explaining the method of determining the travel lane of another vehicle by the luminance distribution in a window. 12 is a flowchart illustrating a detailed flow of another vehicle travel lane determination process performed in the second embodiment.

  Embodiments of an approaching vehicle detection device according to the present invention will be described below with reference to the drawings.

In the first embodiment, the present invention monitors the road on the rear side of the traveling vehicle, and the driver changes the lane when there is an approaching vehicle in a predetermined distance range on the rear side of the adjacent lane. The present invention is applied to an approaching vehicle detection device having a BSW function that outputs a warning and alerts when an intention is indicated.
(Explanation of application scene of approaching vehicle detection device)

  First, the outline of the approaching vehicle detection device 100a (see FIG. 2) and the road environment to which the approaching vehicle detection device 100a is applied will be described with reference to FIGS. 1A and 1B. The approaching vehicle detection device 100a operates during traveling on a road having three or more lanes on one side. Although the approaching vehicle detection device 100a operates regardless of day or night, in particular, by applying the present invention, the approaching vehicle detection device 100a operates more effectively at night.

  FIG. 1A is a diagram illustrating an example of a road 30 to which the approaching vehicle detection device 100a is applied. The road 30 includes a traveling lane 20 on which the vehicle 10 on which the approaching vehicle detection device 100 a is mounted is traveling, an adjacent lane 22 adjacent to the right side of the vehicle 10, and two adjacent lanes 22 adjacent to the adjacent lane 22. It consists of three lanes, a lane 24 (hereinafter referred to as the adjacent lane 24). The left and right ends of the traveling lane 20 are partitioned by lane markers L1 and L2, the left and right ends of the adjacent lane 22 are respectively partitioned by lane markers L2 and L3, and the left and right ends of the adjacent adjacent lane 24 are respectively lane markers L3 and L4. It is partitioned.

  Now, it is assumed that the vehicle 10 is traveling on the road lane 20 of the road 30 in the left direction in FIG. 1A (in the direction of the arrow 10d). At this time, the rear camera 50 installed backward at the rear end of the vehicle 10 captures an imaging range ω including at least the road surface of the traveling lane 20, the adjacent lane 22, and the adjacent adjacent lane 24. Then, the other vehicle 12 (in the direction of the arrow 12d) approaching the vehicle 10 that exists inside the predetermined detection range (window) W1 set on the rear side of the vehicle 10 from the captured image. Detect progress). Then, when the other vehicle 12 approaching the vehicle 10 is detected, and the driver of the vehicle 10 changes the lane to the adjacent lane 22 side even though the other vehicle 12 is approaching. When trying to do so, a warning to be described later is issued to alert.

  Here, the detection range (window) W <b> 1 is set, for example, in a range from 3 m to 30 m behind the vehicle 10 and in a range from 3 m from the right end of the vehicle 10.

  At this time, it is assumed that the other vehicle 12 lights the headlamps 12h and 12h to illuminate the inside of the irradiation range 12r. Further, it is assumed that the other vehicle 14 traveling in the adjacent adjacent lane 24 in the direction of the arrow 14d lights the headlamps 14h and 14h to illuminate the inside of the irradiation range 14r.

  FIG. 1A illustrates only the area on the right rear side of the vehicle 10, but the approaching vehicle detection device 100 a similarly acts on the area on the left rear side of the vehicle 10. That is, as shown in FIG. 1B, when the vehicle 10 is traveling in the traveling lane 20, the rear camera 50 has the traveling lane 20 with the lane markers L 1 and L 2 at the left and right ends, and the lane markers L 5 and L 1 respectively in the left and right directions. The imaging range ω including the adjacent lane 26 as an end is imaged. The other vehicle 16 approaching the vehicle 10 (in the direction of the arrow 16d) is present inside the predetermined detection range (window) W2 set on the rear side of the vehicle 10 from the captured image. Detect progress). When the other vehicle 16 approaching the vehicle 10 is detected, and the driver of the vehicle 10 changes the lane to the adjacent lane 26 side even though the other vehicle 16 is approaching. When trying to do so, a warning is issued and alerted as described later.

  Here, the detection range (window) W2 is set, for example, in a range from 3 m to 30 m behind the vehicle 10 and in a range from 3 m from the left end of the vehicle 10.

At this time, it is assumed that the other vehicle 16 lights the headlamps 16h and 16h to illuminate the inside of the irradiation range 16r. Further, the other vehicle 18 traveling in the direction of the arrow 18d in the two adjacent lanes 28 (hereinafter referred to as the adjacent adjacent lanes 28) having the lane markers L6 and L5 at the left and right ends respectively has the headlamps 18h and 18h. It is assumed that the inside of the irradiation range 18r is illuminated and is illuminated.
(Description of system configuration of approaching vehicle detection device)

  Next, the hardware configuration of the approaching vehicle detection device of this embodiment will be described with reference to FIG. The approaching vehicle detection device 100a according to the present embodiment is mounted on the vehicle 10 and, as shown in FIG. 2, the road surface on the rear side of the vehicle 10 installed rearward near the rear license plate of the vehicle 10 is provided. It has a rear camera 50 that captures an image including it. Further, the ECU 60 includes an ECU 60 that performs recognition processing of an image captured by the rear camera 50 to detect an approaching vehicle and controls an alarm for notifying the presence of the approaching vehicle. The ECU 60 is connected to a CAN bus 80 disposed inside the vehicle 10. The CAN bus 80 is connected to a vehicle speed sensor 82 that detects the vehicle speed v of the vehicle 10 and a winker that detects a driver's turn signal operation. A switch 84 is connected. Further, an indicator 90 that notifies the presence of an approaching vehicle as visual information and a buzzer 92 that notifies the presence of an approaching vehicle as auditory information are connected to the ECU 60.

  The rear camera 50 has an AGC function that automatically changes the gain value according to the ambient brightness. That is, when the surroundings are dark, the gain value is automatically increased to brighten the image, and when the surroundings are bright, the gain value is automatically decreased to darken the image. And the rear camera 50 shall output the gain value at that time with the imaged image.

  Inside the ECU 60, a plurality of software modules for executing necessary processes are mounted. The night determination unit 61 determines whether the ambient environment in which the approaching vehicle detection device 100a is operating is day or night. The vehicle information acquisition unit 62 acquires the vehicle speed v of the vehicle 10 detected by the vehicle speed sensor 82, the operation signal of the turn signal switch 84, and the like via the CAN bus 80. The image recognition processing unit 63 detects a headlamp of a vehicle traveling behind the vehicle 10 and detects an approaching vehicle. The other vehicle travel lane determination unit 64 determines whether the detected other vehicle is traveling in the adjacent lane or in the adjacent adjacent lane. The alarm control unit 65 outputs or suppresses an alarm based on the detection result of the approaching vehicle and the operation of the driver of the vehicle 10. The ECU 60 is also equipped with a software module (not shown in FIG. 2), for example, a module such as an overall control section for controlling the movement of the entire software, and a storage section for storing necessary images and information.

  Next, the functional configuration of the approaching vehicle detection device 100a of this embodiment, particularly the detailed configuration of the image recognition processing unit 63 and the alarm control unit 65, will be described with reference to FIG.

  As shown in FIG. 3, the image recognition processing unit 63 detects the headlamps of other vehicles (12, 14 (FIG. 1A), 16, 18 (FIG. 1B)) from the images captured by the rear camera 50. A headlamp detector 70 is provided. Moreover, it has the overhead image production | generation part 71 which converts the image imaged with the rear camera 50 into the bird's-eye view image which looked down from the front upper side of the vehicle 10. FIG. And it has the window setting part 72 which sets the detection range (window) W1 (FIG. 1A) and W2 (FIG. 1B) of a predetermined size in the predetermined position of the converted bird's-eye view image. Moreover, it has the brightness distribution calculation part 73a which calculates the average value of the brightness | luminance of the pixel corresponding to the set detection range W1, W2 in the bird's-eye view image. The luminance distribution calculating unit 73a is connected to a moving average calculating unit 74 that calculates a moving average value of average luminance inside the detection ranges W1 and W2.

  The image recognition processing unit 63 further includes an image alignment unit 75 that aligns the overhead images obtained at different times in the overhead image generation unit 71. Moreover, it has the difference image generation part 76 which performs the frame difference of the bird's-eye view image aligned, and produces | generates a difference image. And it has the relative speed calculation part 77 which calculates the relative speed of the vehicle 10 and a moving object based on the result of having performed the frame difference. Moreover, it has the three-dimensional object determination part 78 which determines whether the moving object reflected inside the detection range W1, W2 is another vehicle (12, 14 (FIG. 1A), 16, 18 (FIG. 1B)).

As shown in FIG. 3, the alarm control unit 65 operates when the driver of the vehicle 10 operates the turn signal to change the lane when the approaching vehicle is detected and the approaching vehicle is detected. An alarm output unit 66 that outputs an alarm when the display is performed, and an alarm suppression unit 67 that suppresses the output of the alarm when an approaching vehicle is detected in the adjacent lane.
(Description of the operation of the headlamp detector)

  Next, the operation of the headlamp detection unit 70 of the approaching vehicle detection device 100a will be described with reference to FIG.

  The greatest feature of the approaching vehicle detection device 100a shown in the first embodiment is that when detecting other vehicles (12, 14 (FIG. 1A), 16, 18 (FIG. 1B)) approaching the vehicle 10 from the rear. It is possible to determine whether another vehicle is traveling in the adjacent lane (22 in FIG. 1A, 26 in FIG. 1B) or in the adjacent adjacent lane (24 in FIG. 1A, 28 in FIG. 1B). And the approaching vehicle detection apparatus 100a exhibits the effect especially at night.

  At night, as shown in FIG. 4, the original image I (t) picked up at time t by the rear camera 50 (FIG. 3) becomes a dark image because the overall luminance is low. Therefore, it is very difficult to recognize the body of another vehicle from the original image I (t).

  Therefore, the approaching vehicle detection device 100a recognizes the presence of the other vehicle by detecting the headlamp image of the other vehicle. At this time, among other vehicles, there is another vehicle (corresponding to the headlamp image 120 in FIG. 4) traveling in the same travel lane 20 (FIG. 4) as the vehicle 10. Since the other vehicle traveling in the same traveling lane as the vehicle 10 is not a detection target of the approaching vehicle detection device 100a in this way, a headlight image (for example, FIG. It is necessary to detect the headlamp image 122).

  Therefore, in the approaching vehicle detection device 100a, as shown in FIG. 4, headlight detection regions R1, R2 (predetermined regions) are set on the left and right sides so as to avoid the traveling lane 20. And the area | region which has the high brightness | luminance more than predetermined brightness | luminance is detected in the set headlamp detection area | regions R1 and R2, and when the applicable area | region is detected, it shall be a headlamp image of another vehicle. . That is, in the case of the original image I (t) shown in FIG. 4, the headlamp image is not detected inside the headlamp detection region R1, and the headlamp image 122 is detected inside the headlamp detection region R2. The The headlamp detection areas R1 and R2 set here are set to a range where at least a headlamp image of another vehicle existing in the adjacent lane can be detected. The specific dimensions may be determined appropriately by conducting experiments or the like in advance.

  In the daytime, the vehicle body of the other vehicle can be visually recognized in the image captured by the rear camera 50 (FIG. 3), and thus it is not necessary to perform the above-described headlamp image detection process. Therefore, in the approaching vehicle detection device 100a, it is determined whether or not the nighttime determination unit 61 (FIG. 3) performs the headlamp image detection process.

Specifically, the night determination unit 61 receives a gain value output when the rear camera 50 performs imaging, and determines that the surroundings are dark when the gain value is equal to or greater than a predetermined gain. It is assumed that a lamp image detection process is performed.
(Explanation of functions of luminance distribution calculation unit and moving average calculation unit)

  Next, operations of the luminance distribution calculation unit 73a and the moving average calculation unit 74 of the approaching vehicle detection device 100a will be described with reference to FIGS. 5 and 6A. Note that the image captured by the rear camera 50 is converted into an overhead image in which the vehicle 10 is viewed from directly above in the overhead image generation unit 71, and the processing described below is performed on this overhead image. And

  The luminance distribution calculating unit 73a calculates average values (luminance evaluation values) of the luminance within the detection ranges (windows) W1 and W2 set by the window setting unit 72, respectively. Here, as shown in FIG. 5, the detection range W <b> 1 is set on the overhead image from the left end of the vehicle 10 to the width wa in the left direction. Further, it is set over a length wb from a position behind the rear end of the vehicle 10 by a distance d. As described above, for example, wa = 3m, d = 3m, and wb = 27m are set as specific values of the width wa, the distance d, and the length wb.

  Next, the operation of calculating the luminance distribution inside the detection ranges (windows) W1 and W2 will be described with reference to FIG. 6A.

  As shown in FIG. 6A, other vehicles 14 and 16 approach from the rear of the vehicle 10, and the other vehicle 14 travels in the adjacent lane 24, and the other vehicle 16 travels in the adjacent lane 26. It shall be. The other vehicle 14 lights the headlamp, and its irradiation range is 14r. The other vehicle 16 also lights the headlamp, and its irradiation range is 16r.

  At this time, as can be seen from FIG. 6A, the headlamp illumination range 14r of the other vehicle 14 partially overlaps the detection range (window) W1. On the other hand, the headlamp illumination range 16r of the other vehicle 16 almost overlaps the detection range (window) W2. Therefore, the travel lanes of the other vehicles 14 and 16 can be estimated based on the average brightness inside the detection ranges W1 and W2.

The luminance distribution calculation unit 73a calculates the average luminance (luminance evaluation value) of the pixels in the area corresponding to the detection ranges (windows) W1 and W2, respectively, with respect to the overhead image. The average luminance is calculated as a value obtained by dividing the sum of the pixel values in the areas corresponding to the detection ranges W1 and W2 by the total number of pixels in the detection ranges W1 and W2, respectively, in the overhead image. Here, the average luminance of the pixels inside the detection range W1 at time t is represented by B _W1 (t), and the average luminance of the pixels inside the detection range W2 is represented by B _W2 (t).

The brightness of the road surface of the road 30 has a high value at a place where the road illumination is bright, even at night. Therefore, the average luminances B _W1 (t) and B _W2 (t) calculated by the luminance distribution calculation unit 73a are simply compared with the fixed threshold value, and the detection ranges W1 and W2 and the irradiation of the headlamps of other vehicles are performed. It is difficult to determine the overlapping state with the ranges 14r and 16r.

Therefore, the moving average calculation unit 74 calculates the moving average values B _thW1 (t) and B _thW2 (t) of the average luminances B _W1 (t) and B _W2 (t), respectively, and this moving average value B _thW1 (t ) And B _thW2 (t) are used as threshold values, respectively, and the overlapping state between the detection ranges W1 and W2 and the irradiation ranges 14r and 16r of the headlamps of other vehicles is determined. Details will be described later.

It should be noted that the time length for calculating the moving average value may be set to an appropriate value by conducting an experiment or the like. However, in order to increase the determination accuracy, the headlamps of other vehicles are irradiated within the detection ranges W1 and W2. It is desirable to include a period that is not done.
(Description of the operation of the other vehicle travel lane determination unit)

  Next, the operation of the other vehicle travel lane determination unit 64 will be described with reference to FIG. 6B.

As shown in FIG. 6B, the other vehicle travel lane determination unit 64 uses the average brightness B _W1 (t) and B _W2 (t) calculated by the brightness distribution calculation unit 73a to calculate the movement calculated by the moving average calculation unit 74. The average values B _thW1 (t) and B _thW2 (t) are respectively compared.

When B _W1 (t) ≧ B _thW1 (t), it is determined that the other vehicle 14 is traveling in the adjacent lane 22. When B _W1 (t) <B _thW1 (t), it is determined that the other vehicle 14 is traveling in the adjacent adjacent lane 24.

Similarly, when B _W2 (t) ≧ B _thW2 (t), it is determined that the other vehicle 16 is traveling in the adjacent lane 26. When B _W2 (t) <B _thW2 (t), it is determined that the other vehicle 16 is traveling in the adjacent adjacent lane 28.

Thus, by using the moving average values B _thW1 (t) and B _thW2 (t) as threshold values, when the road surface brightness of the road 30 (FIG. 6A) fluctuates, for example, when the road surface is bright, the moving average When the values B _thW1 (t) and B _thW2 (t) become large and the road surface is dark, the moving average values B _thW1 (t) and B _thW2 (t) become small, so the brightness of the road surface by the headlamps of other vehicles. A change can be reliably detected.
(Description of the action of the three-dimensional object detection unit)

  The approaching vehicle detection device 100a processes an image captured by the rear camera 50 to detect an approaching vehicle. An overview of image processing performed at that time will be described. The series of processing performed here is applied to a vehicle equipped with the BSW function and is put into practical use, and therefore, the description with reference to the drawings is omitted and only the text is described.

  First, two images captured at different times separated by the rear camera 50 at a time interval Δt are converted into overhead images by the overhead image generation unit 71, respectively.

  Next, in the image alignment unit 75, among the two generated overhead images, the temporal overhead image is transformed and superimposed with the temporally new overhead image. This deformation is performed by using the vehicle speed v of the vehicle 10 during the time interval Δt acquired by the vehicle information acquisition unit 62 and the steering angle of the vehicle 10 during the time interval Δt not shown in FIG. It is performed so as to correspond to the behavior of the vehicle 10 during the period. In other words, the overhead view image is translated and rotated so as to correspond to the moving direction and the moving amount of the vehicle 10 in the time interval Δt, and the alignment is performed. By this alignment, the road surface areas of the roads of the two overhead images completely overlap.

  Thereafter, the difference image generation unit 76 performs an inter-frame difference between the two overhead images that have been aligned. At this time, the past overhead image is subtracted from the temporally new overhead image. By this difference calculation, a region where a change in luminance occurs during the time interval Δt, that is, a region where a moving object exists is detected.

  Next, the relative speed calculation unit 77 calculates the relative speed between the vehicle 10 and the moving object by obtaining the size of the area where the moving object exists (length in the front-rear direction of the vehicle 10).

Then, the three-dimensional object determination unit 78 confirms that the moving object is another vehicle. In this process, for example, an area recognized as having a moving object on the overhead image is inversely transformed onto the original image I (t) captured by the rear camera 50, and the object exists in the inversely transformed area. This is done by determining whether or not. The determination of whether or not an object exists may be performed by template matching by applying a template imitating the shape of the vehicle in the vicinity of the inversely transformed region. Alternatively, the luminance of each pixel of the original image I (t) is projected in the vertical direction and the horizontal direction, and there is a change point in the portion corresponding to the inversely transformed area in the obtained luminance projection curve. After confirming this, it may be determined that another vehicle exists.
(Description of the action of the alarm control unit)

  The alarm control unit 65 controls the alarm output for the approaching vehicle based on the detection result of the other vehicle, the vehicle speed v of the vehicle 10 and the operation state of the turn signal switch 84.

  Specifically, when the vehicle speed v of the vehicle 10 is equal to or higher than a predetermined vehicle speed (for example, 30 km / h), when an approaching vehicle is detected on the adjacent lane 22 or the adjacent lane 26 (FIG. 6A), the alarm output unit 66. The indicator 90 is turned on by the action of.

  Further, when the driver of the vehicle 10 operates the turn signal switch 84 without noticing the lighting of the indicator 90 and issues a turn signal to the adjacent lane 22 or the adjacent lane 26 side where the approaching vehicle is detected, the buzzer 92 is further turned on. Raise a ring and call attention.

  When an approaching vehicle is detected on the adjacent adjacent lane 24 or the adjacent adjacent lane 28 (FIG. 6A), the above-described control is not executed. That is, neither the indicator 90 is lit nor the buzzer 92 is sounded.

Alternatively, when an approaching vehicle is detected on the adjacent adjacent lane 24 or the adjacent adjacent lane 28 (FIG. 6A), the output of the alarm 90 is suppressed for a certain time by the operation of the alarm suppression unit 67, and the indicator 90 is turned on. Alternatively, the buzzer 92 may not be sounded.
(Description of the flow of processing performed by the approaching vehicle detection device)

  Next, the overall flow of processing performed by the approaching vehicle detection device 100a will be described with reference to FIG. In addition, FIG. 3 is referred for description of a structure part.

  (Step S10) The vehicle information acquisition unit 62 acquires vehicle information. Specifically, the vehicle speed v of the vehicle 10 obtained by the vehicle speed sensor 82 and the steering angle of the vehicle 10 not shown in FIG. 3 are acquired.

  (Step S <b> 12) An image behind the vehicle 10 captured by the rear camera 50 is input to the image recognition processing unit 63.

  (Step S14) The night determination unit 61 acquires the gain value output from the rear camera 50.

  (Step S16) An overhead image is generated from the image input in the overhead image generation unit 71.

  (Step S18) The image alignment unit 75 aligns two overhead images obtained at different times.

  (Step S20) The difference image generation unit 76 performs frame difference between the two overhead images that have been aligned, and generates a difference image.

  (Step S22) The relative speed calculation unit 77 calculates the relative speed of the area representing the moving object detected from the difference image.

  (Step S24) The night determination unit 61 performs day / night determination. When it is determined that it is noon, the process proceeds to step S30, and when it is determined that it is night, the process proceeds to step S26.

  (Step S26) The headlamp detection unit 70 detects the headlamp of another vehicle.

  (Step S28) The other vehicle travel lane determination unit 64 performs another vehicle travel lane determination process in which the other vehicle travels and determines the lane. The detailed flow of the other vehicle travel lane determination process will be described later.

  (Step S30) The three-dimensional object determination unit 78 determines whether there is a three-dimensional object (another vehicle) behind.

  (Step S32) The alarm control unit 65 performs predetermined alarm control. The outline of the alarm control is as described above.

  (Step S34) The vehicle speed v of the vehicle 10 acquired by the vehicle information acquisition unit is monitored to determine whether or not the vehicle speed v of the vehicle 10 has fallen below a predetermined vehicle speed. When the vehicle speed falls below a predetermined vehicle speed (for example, 30 km / h), the processing of FIG.

In the flowchart of FIG. 7, the day / night determination is performed in step S24, but the day / night determination may be performed after step S14.
(Description of the flow of other vehicle travel lane determination processing)

  Next, the flow of the other vehicle travel lane determination process will be described with reference to FIG. In addition, FIG. 3 is referred for description of a structure part.

  (Step S50) The window setting unit 72 sets detection ranges (windows) W1 and W2 on the overhead view image.

(Step S52) The luminance distribution calculation unit 73a calculates average luminances B _W1 (t) and B _W2 (t) inside the detection ranges (windows) W1 and W2, respectively.

(Step S54) The moving average calculation unit 74 _updates the moving average values B _thW1 (t) and B _thW2 (t) of the average luminance inside the detection ranges (windows) W1 and W2.

(Step S56) In the other vehicle travel lane determining unit 64, whether or not the average luminance B _W1 (t) is equal to or higher than the moving average value B _thW1 (t), and the average luminance B _W2 (t) is the moving average value B _thW2 (t ) Determine whether or not. When the average luminance is equal to or higher than the moving average value, the process proceeds to step S60, and otherwise, the process proceeds to step S58.

  (Step S58) It is determined that the other vehicle is traveling in the adjacent adjacent lane, and the process returns to the main routine (FIG. 7).

  (Step S60) It is determined that the other vehicle is traveling in the adjacent lane, and the process returns to the main routine (FIG. 7).

Next, a second embodiment of the approaching vehicle detection device according to the present invention will be described with reference to the drawings. In the second embodiment, as in the first embodiment, the present invention is applied to an approaching vehicle detection device having a BSW function.
(Description of system configuration of approaching vehicle detection device)

  First, the functional configuration of the approaching vehicle detection device 100b of this embodiment will be described with reference to FIG. As shown in FIG. 9, the basic functional configuration is the same as the functional configuration of the first embodiment (FIG. 3), but the luminance distributions within the detection ranges (windows) W1, W2 set by the window setting unit 72 are as follows. Only the calculation method is different. That is, instead of the luminance distribution calculation unit 73a and the moving average calculation unit 74 included in the first embodiment (FIG. 3), the second embodiment includes a luminance projection distribution calculation unit 73b.

The brightness projection distribution calculation unit 73b is configured to extend along the direction (the left-right direction of the vehicle 10) crossing the detection ranges W1 and W2 with respect to the detection ranges (windows) W1 and W2 set by the window setting unit 72, respectively. It has a function of calculating the brightness projection value by accumulating the brightness of the corresponding pixels of the overhead image.
(Description of the operation of the luminance projection distribution calculation unit)

  Next, the operation of the luminance projection distribution calculation unit 73b of the approaching vehicle detection device 100b will be described.

As described above, the brightness projection distribution calculation unit 73b accumulates the brightness of the corresponding pixels of the overhead image along the direction (the left-right direction of the vehicle 10) that crosses the detection ranges (windows) W1 and W2 (FIG. 6A). Then, the brightness projection values Bp _W1 and Bp _W2 are calculated, and the brightness projection distributions Bp _W1 (Z) and Bp _W2 (Z), which are the distribution curves of the brightness projection values in the front-rear direction of the detection ranges W1 and W2, are generated. To do. Here, Z is a coordinate that specifies the position in the front-rear direction of the detection ranges W1 and W2, and is a coordinate that represents the rear position of the vehicle 10.

FIG. 10 is a diagram illustrating an example of the luminance projection distributions Bp _W1 (Z) and Bp _W2 (Z) generated when the vehicle 10 is in the state of FIG. 6A.

As can be seen from FIG. 10, when the other vehicle 16 is in the adjacent lane 26, the luminance projection distribution Bp _W2 (Z) is a substantially uniform distribution over the entire detection range W2. On the other hand, when the other vehicle 14 is in the adjacent lane 24, the luminance projection distribution Bp _W1 (Z) is a distribution having a luminance projection value protruding at a specific position in the detection range W1. In the example of FIG. 10, the brightness projection value exhibits a high value at the position closest to the vehicle 10 in the detection range W2.

Therefore, for example, when the standard deviation (luminance evaluation value) σ calculated from the luminance projection distributions Bp _W1 (Z) and Bp _W2 (Z) is equal to or larger than a predetermined standard deviation, the other vehicle is traveling in the adjacent lane. Can be determined. Further, when the standard deviation (luminance evaluation value) σ is less than the predetermined standard deviation, it can be determined that the other vehicle is traveling in the adjacent adjacent lane. Note that an appropriate value is set as the predetermined standard deviation serving as the threshold value based on experiments or the like.
(Description of Flow of Other Vehicle Traveling Lane Determination Processing in Example 2)

  The overall flow of the process performed by the approaching vehicle detection device 100b is substantially the same as the flow of the first embodiment shown in FIG. 7, and only the method of the other vehicle travel lane determination process (step S28 in FIG. 7) is different. . Therefore, illustration of a flowchart showing the overall flow of processing performed in the second embodiment is omitted.

  Hereinafter, the flow of the other vehicle travel lane determination process performed by the approaching vehicle detection device 100b will be described with reference to FIG. In addition, FIG. 9 is referred for description of a component part.

  (Step S80) The window setting unit 72 sets detection ranges (windows) W1, W2 on the overhead image.

(Step S82) The luminance projection distribution calculation unit 73b calculates the luminance projection distributions Bp _W1 (Z) and Bp _W2 (Z) inside the detection ranges (windows) W1 and W2, respectively.

(Step S84) The other vehicle travel lane determination unit 64 determines whether or not the standard deviation σ of the luminance projection distributions Bp _W1 (Z) and Bp _W2 (Z) is equal to or greater than a predetermined standard deviation. When it is greater than or equal to the predetermined standard deviation, the process proceeds to step S88, and otherwise, the process proceeds to step S86.

  (Step S86) It is determined that the other vehicle is traveling in the adjacent adjacent lane, and the process returns to the main routine (not shown).

  (Step S88) It is determined that the other vehicle is traveling in the adjacent lane, and the process returns to the main routine (not shown).

When the other vehicle 14 is in the adjacent adjacent lane 24, the protruding position of the brightness projection distribution Bp _W1 (Z) changes with time according to the relative speed between the vehicle 10 and the other vehicle 14. Therefore, the protruding position of the luminance projection distribution Bp _W1 (Z) can also be used as the luminance evaluation value. That is, the standard deviation σ is less than a predetermined standard deviation of the luminance projection distribution Bp W1 _(Z), yet, when the luminance projection distribution Bp W1 of projecting position of the _(Z) is moving as time, other vehicles 14 The determination accuracy can be improved by determining that the vehicle is traveling in the adjacent lane.

As described above, the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above includes the road surface on the rear side of the vehicle 10 imaged by the rear camera 50 (imaging unit). When the headlamp detection unit 70 detects the headlamp of another vehicle from the original image I (t), the luminance distribution calculation unit 73a performs the overhead view image obtained by performing the overhead conversion of the original image I (t). The average luminance B _W1 (t) (B _W2 (t)) (luminance evaluation value) inside the detection range W1 (W2) (window) of a predetermined size set by the window setting unit 72 is calculated, and the other vehicle Based on the average luminance B _W1 (t) (B _W2 (t)) (luminance evaluation value), the traveling lane determination unit 64 travels in the adjacent lane 22 (26) of the lane in which the vehicle 10 is traveling. Whether the vehicle is driving in the adjacent lane 24 (28) Because, the approaching vehicle traveling the next adjacent lane 24 (28), there is no erroneously detected as the approaching vehicle traveling on the next lane 22 (26).

Further, according to the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above, the luminance distribution calculation unit 73a uses the average luminance within the detection range W1 (W2) (window) as the luminance evaluation value. B _W1 (t) (B _W2 (t)) is calculated, and the other vehicle travel lane determination unit 64 determines that the average luminance B _W1 (t) (B _W2 (t)) is the moving average value B _thW1 (t) ( If B _thW2 (t)) (predetermined average value) is greater, it is determined that the other vehicle is traveling in the adjacent lane 22 (26). Otherwise, the other vehicle is adjacent to the adjacent lane 24 (28). Since it is determined that the vehicle is traveling, it is possible to reliably determine the traveling lane of the other vehicle in a short time by a simple calculation process.

According to the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above, the average luminance B _W1 (t) (B _W2 (t)) for determining the traveling lane of the other vehicle Since the threshold value (predetermined average value) is the moving average value B _thW1 (t) (B _thW2 (t)) of the average luminance B _W1 (t) (B _W2 (t)), the brightness of the road surface of the road 30 Even when the vehicle speed fluctuates, it is possible to reliably detect a change in the brightness of the road surface due to the headlamps of other vehicles.

  Furthermore, according to the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above, the window setting unit 72 is an overhead view of an image captured by the rear camera 50 (imaging unit) from directly above. Since it is converted into an image and at least one detection range W1 (W2) (window) of a predetermined size is set in this overhead image, the detection range W1 (W2) (window) of a predetermined size is easily and reliably set can do. In particular, even if the mounting position of the rear camera 50 is different for each vehicle, it is only necessary to set a window having the same shape at the same position on the overhead image as long as the conversion to the overhead image can be performed. There is no need to prepare different programs.

  Further, according to the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above, the headlamp detection unit 70 includes the vehicle 10 in the image captured by the rear camera 50 (imaging unit). The headlamps of other vehicles are detected within the headlamp detection areas R1 and R2 (predetermined areas) set in a range where the headlamps of other vehicles traveling in the adjacent lane or the adjacent adjacent lane are observed. Therefore, the necessary headlamps can be reliably detected in a short time by narrowing down the detection range of the headlamps.

  And according to the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above, the other vehicle travel lane determination unit 64 determines that the other vehicle is traveling in the adjacent adjacent lane. In addition, since it is determined that there is no other vehicle on the adjacent adjacent lane side, it is possible to reliably prevent false alarms.

  Furthermore, according to the approaching vehicle detection device 100a according to the first embodiment of the present invention configured as described above, when the vehicle 10 is determined that the other vehicle is traveling in the adjacent lane, the vehicle 10 Since it has the alarm suppression part 67 which suppresses the alarm output at the time of trying to change the lane to the adjacent lane side, it is possible to reliably prevent false alarms when other vehicles are traveling in the adjacent lane. Can do.

Further, according to the approaching vehicle detection device 100b according to the second embodiment of the present invention configured as described above, the luminance evaluation value is the luminance value in the detection range W1 (W2) (window), and the vehicle 10 is in the horizontal direction. When the standard deviation σ of the luminance projection distributions Bp _W1 (Z) (Bp _W2 (Z)) accumulated along, and the standard deviation σ is larger than the predetermined standard deviation Because it is determined that the other vehicle is traveling in the adjacent lane and the other vehicle is traveling in the adjacent lane when it is other than that, the other vehicle is reliably determined in a short time by a simple calculation process. The traveling lane of the vehicle can be determined.

  In the first and second embodiments, the detection ranges W1 and W2 are both set. However, in FIG. 5, the detection is performed when the vehicle 10 travels to the right of the road 30 (adjacent lane 22, adjacent lane 24). Only the range W2 may be set and only the detection range W1 may be set when the vehicle 10 is traveling to the left of the road 30 (adjacent lane 26, adjacent adjacent 28). Whether the vehicle 10 is traveling on the right side or the left side of the road 30 is not described in the first and second embodiments. For example, the lane marker drawn on the lane edge of the road is detected and the arrangement thereof is analyzed. It can be judged by.

  Further, in the first and second embodiments, an example is shown in which processing is performed using only the image captured by the rear camera 50. However, in addition to the rear camera 50, the left and right door mirrors built in the left and right door mirrors of the vehicle 10 are also included. It is also possible to perform the same processing using an image obtained by capturing a wider range using a rear side camera.

  As mentioned above, although the Example of this invention was explained in full detail with drawing, since an Example is only an illustration of this invention, this invention is not limited only to the structure of an Example. Of course, changes in design and the like within a range not departing from the gist are included in the present invention.

10 .... Vehicle 50 ...... Rear camera 61 ... Night determination unit 62 ... Vehicle information acquisition unit 63 ... Image recognition processing unit 64 ... Other vehicle travel lane determination unit 65 ... Alarm control unit 66 ... Alarm output unit 67 ... Alarm suppression unit 70 ... Headlight detection unit 71 ... Overhead image generation unit 72 ... Window setting Unit 73a ... Luminance distribution calculator 74 ... Moving average calculator 75 ... Image registration unit 76 ... Difference image generator 77 ... Relative velocity calculator 78 ... Three-dimensional object determination unit 82... Vehicle speed sensor 84... Turn signal switch 90... Indicator 92.

Claims (8)

An approaching vehicle detection device that detects an adjacent lane of a lane in which the vehicle is traveling from a running vehicle, detects other vehicles approaching from behind by turning on a headlamp,
An imaging unit that captures an image including a road surface on a rear side of the vehicle;
A window setting unit for setting at least one window of a predetermined size having a predetermined positional relationship with the vehicle at a position corresponding to a rear side region of the vehicle in the image;
A headlamp detector that determines whether or not a headlamp of another vehicle is detected in a predetermined region of the image;
Luminance distribution calculation for calculating a luminance evaluation value in each window based on a luminance value of an area corresponding to each window of the image when the headlamp detection unit detects a headlamp of another vehicle And
Based on the luminance evaluation value in each window, the other vehicle exists in an adjacent lane adjacent to the lane in which the vehicle is traveling, or is adjacent to the far side of the vehicle with respect to the adjacent lane. An approaching vehicle detection device comprising: another vehicle traveling lane determination unit that determines whether the vehicle is in an adjacent adjacent lane. The luminance evaluation value is an average value of luminance values of regions corresponding to the windows,
The other vehicle travel lane determining unit determines that the other vehicle is traveling in an adjacent lane when the average luminance is greater than a predetermined average value, and otherwise, the other vehicle is in an adjacent adjacent lane. The approaching vehicle detection device according to claim 1, wherein it is determined that the vehicle is traveling.   The approaching vehicle detection device according to claim 2, wherein the predetermined average value is a moving average value of luminance values of regions corresponding to the windows. The luminance evaluation value is a standard deviation of the luminance projection distribution obtained by accumulating the luminance value in each window along the left-right direction of the vehicle,
The other vehicle travel lane determining unit determines that the other vehicle is traveling in an adjacent lane when the standard deviation is larger than a predetermined standard deviation, and when the other vehicle is other than that, the other vehicle is The approaching vehicle detection device according to claim 1, wherein the approaching vehicle detection device determines that the vehicle is traveling in an adjacent adjacent lane.   The window setting unit converts an image captured by the image capturing unit into an overhead image viewed from directly above, and sets at least one window of a predetermined size in the overhead image. The approaching vehicle detection device according to any one of claims 1 to 4.   The said predetermined area is set to the range in which the headlamp of the other vehicle which is drive | working the adjacent lane of the said vehicle or the adjacent adjacent lane in the said image is observed. Item 6. The approaching vehicle detection device according to any one of item 5.   The said other vehicle travel lane determination part judges that the other vehicle does not exist in the said adjacent adjacent lane side, when it determines with the said other vehicle drive | working the adjacent adjacent lane. The approaching vehicle detection device according to any one of claims 1 to 6.   The vehicle has an alarm suppression unit that suppresses an alarm output when the vehicle tries to change the lane to the adjacent adjacent lane when it is determined that the other vehicle is traveling in the adjacent adjacent lane. The approaching vehicle detection device according to any one of claims 1 to 6, wherein: