JP2003084064A - Device and method for recognizing vehicle in front side - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate reflection from a vehicle other than a vehicle in a front side, a road side matter and the like to conduct highly precise recognition, by integrating vehicle recognition by an image sensor, when the vehicle is recognized using a laser radar. SOLUTION: A vehicle point proposal group is coordinate-transformed into a camera coordinate system of a CCD camera, by a CPU 3, using as the vehicle point proposal group a reflection point group existing within a spreading range of a substantial vehicle width in a position of a substantial equal distance, based on positions of respective reflection points specified by a laser radar module 1, so as to be collated with an extracted rectangular area extracted by a camera module 2, and the vehicle point proposal group is judged as the vehicle in the front side when the vehicle point proposal group after coordinate- transformed is substantially consistent with the rectangular area, by the computer 3.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a front vehicle recognizing device and a recognizing method for recognizing a vehicle traveling in front of an own vehicle.

[0002]

2. Description of the Related Art In recent years, a vehicle equipped with a more advanced and more comfortable driving support system, such as a follow-up running function and a steering assist function on a highway, has been proposed. There is a recognition device that recognizes the vehicle.

In a conventional recognition device using this kind of scan laser radar, the distance to the reflection point can be detected by measuring the time from the irradiation of the laser beam to the observation of the reflected light, and the scanning mechanism is provided. By providing 10
If the vehicle has a horizontal field of view of several degrees and is a long-distance vehicle,
As shown in Fig. 1, two reflections from the reflectors (reflectors) mounted on both ends of the vehicle rear end face can be observed, and many reflections from the body of the vehicle can be observed at medium and short distances. The numbers range from 3 to more. FIG. 11 (a)
Represents the situation in which the laser light is emitted from the laser radar installed in the own vehicle, and ● in the figure (b) indicates the reflection point. In the figure, ZR is the traveling direction of the own vehicle and XR is orthogonal to this. It is the direction to do.

However, in this case, there is no constant reflection from the same part of the vehicle body, and the direction, shape,
The reflection position changes every moment due to the positional relationship and the distribution is not uniform.

Therefore, there has been proposed a method of recognizing a preceding vehicle by clustering the obtained reflection point group by a fuzzy method or the like. As a specific example thereof, Preprints 931 No. 93017
19, pp53-56 (1993-5) ("Development of preceding vehicle recognition algorithm by laser radar"), or Japanese Patent Application Laid-Open No. 6-30
There is a method described in Japanese Patent No. 9600.

[0006]

However, these conventional methods have almost no problem when the distance between vehicles or the distance between vehicles and a roadside structure is wide like an expressway. If you try to expand to, the following problems will occur. That is, as the first problem, the reflection from the traveling vehicle or the parked vehicle in the vicinity of the front vehicle is included in the reflection from the front vehicle and processed, and an error occurs in the distance measurement value of the front vehicle. The problem is that the reflection from the roadside object is included in the reflection from the vehicle in front and processed.
As a third problem that an error occurs in the distance measurement value of the vehicle ahead, there is a problem that detection of an interrupting vehicle or the like at a short distance is delayed because the horizontal field of view of the radar is generally narrow.

Further, in the case of a laser radar, when no vehicle is present, reflection from a roadside structure such as a reflector of a guardrail or a signboard is also observed, and these are erroneously recognized as a parked vehicle or a forward vehicle, or they are adjacent to each other. It may be determined that the vehicle in front of the lane is a vehicle traveling in the same lane as the observation vehicle, or the laser light may be attenuated by the water in front of the road splashing by the vehicle in front, making it impossible to recognize the vehicle in front. Is a problem peculiar to a laser radar or another problem requiring a very high level of recognition processing, and is not a problem to be solved by the present invention this time.

Therefore, in the present invention, when the vehicle recognition is performed by using the laser radar, the vehicle recognition by the image sensor is combined to eliminate the reflection from the vehicle other than the vehicle in front, the roadside object, etc. The purpose is to be able to perform high recognition.

[0009]

In order to achieve the above object, a recognition device for a front vehicle according to the present invention irradiates a front side of a vehicle with a laser beam while scanning the laser beam in a horizontal direction and reflects the laser beam from a reflection point. A laser radar that receives light and specifies the positions of a plurality of the reflection points, an image sensor that images the front of the vehicle, and a predetermined rectangular area including the rear end face image of the vehicle ahead is extracted from the image captured by the image sensor. The image processing unit, the position of each of the reflection points specified by the laser radar and the rectangular area by the image processing unit are collated, and the reflection point at the position corresponding to the rectangular area is reflected on the front vehicle. And a recognition unit for recognizing the preceding vehicle.

According to this structure, the position of each reflection point specified by the laser radar is compared with the rectangular area by the image processing unit, and the reflection point at the position corresponding to the rectangular area is reflected on the vehicle ahead. It is judged as a point. At this time, the laser radar excels in distance measurement, and the image sensor can accurately distinguish between the vehicle and the other. Therefore, by combining the vehicle recognition by the image sensor with the vehicle recognition by the laser radar, it is possible to use a vehicle other than the front vehicle. It is possible to prevent the reflection from the vehicle or the roadside object from being included in the reflection from the front vehicle, and it is possible to recognize the front vehicle with high accuracy.

Further, in the recognition device for a front vehicle according to the present invention, the recognition unit is located at substantially equidistant positions on the basis of the positions of the respective reflection points specified by the laser radar and within a widened range of the vehicle width. The existing reflection point group is used as a vehicle candidate point group, the vehicle candidate point group is coordinate-converted into the coordinate system of the image sensor and collated with the rectangular area, and the vehicle candidate point group after coordinate conversion is It is characterized in that the vehicle candidate point group is determined to be a forward vehicle when the vehicle candidate point group substantially coincides with the rectangular area.

According to this structure, the collation unit
Based on the position of each reflection point specified by the laser radar, the reflection point group existing in the vehicle width spread range at almost equidistant positions is set as the vehicle candidate point group, and this vehicle candidate point group is detected by the image sensor. When the coordinate-converted vehicle candidate point group is substantially matched with the rectangular area, the vehicle candidate point group is determined to be a forward vehicle by the integration unit. Therefore, it is possible to reliably prevent the reflection from the vehicle other than the front vehicle, the reflection from the roadside object, and the like, as in the conventional case, to contribute to the improvement of the recognition accuracy of the front vehicle. .

Further, in the front vehicle recognition device according to the present invention, the integration unit predicts the motion of the vehicle candidate point group judged to be a front vehicle by a Kalman filter, and uses it for the next verification process by the verification unit. It is characterized by that.

According to such a configuration, the movement of the vehicle candidate point group determined to be the forward vehicle is predicted and used in the next matching process, so that more accurate vehicle recognition can be realized.

Further, in the front vehicle recognition apparatus according to the present invention, the image processing unit sets a predetermined gazing area for an image picked up by the image sensor, and from the gray level of each pixel in the gazing area to the X direction. Is calculated, and the maximum value is derived to calculate the product of the edge histogram in the Y direction and the edge histogram in both directions orthogonal thereto, and the maximum point is derived to extract the rectangular area as a vehicle candidate. .

According to this structure, when a rectangular area as a vehicle candidate is extracted from the edge histogram,
Other than the vehicle can be reliably excluded, and a rectangular area as a vehicle candidate can be accurately extracted.

Further, in the method for recognizing a front vehicle according to the present invention, a laser radar irradiates the front side of the vehicle with laser light while scanning the same in the horizontal direction, receives reflected light from a reflection point, and receives a plurality of the reflected lights. Identifying the position of the points,
The step of extracting a predetermined rectangular area including the rear end surface image of the front vehicle from the imaged image obtained by imaging the front of the own vehicle by the image sensor, the position of each reflection point specified by the laser radar, and the rectangle The method is characterized by including a step of matching the area and a step of recognizing the front vehicle by determining the reflection point at a position corresponding to the rectangular area as a reflection point on the front vehicle.

According to this structure, since the vehicle recognition by the image sensor is combined with the vehicle recognition by the laser radar, the reflection from the vehicle other than the front vehicle or the roadside object is reflected from the front vehicle as in the conventional case. It is possible to prevent the vehicle in front from being included in the vehicle and to recognize the vehicle in front with high accuracy.

Further, in the method for recognizing a forward vehicle according to the present invention, the reflection points existing at substantially equidistant positions within a substantially widened range of the vehicle width based on the positions of the reflection points specified by the laser radar. A group of vehicle candidate points, the step of coordinate-converting the vehicle candidate point group into the coordinate system of the image sensor and collating with the rectangular area; and the vehicle candidate point group after coordinate conversion substantially coincides with the rectangular area. And a step of determining the vehicle candidate point group as a forward vehicle when the vehicle is driven.

According to such a configuration, the reflection point groups existing in the substantially wide range of the vehicle width at substantially equidistant positions are regarded as vehicle candidate point groups based on the positions of the respective reflection points specified by the laser radar. Then, the vehicle candidate point group is coordinate-converted into the coordinate system of the image sensor and collated with the rectangular area, and when the vehicle candidate point group after the coordinate conversion substantially coincides with the rectangular area, the vehicle candidate point group becomes the front vehicle. Since it is determined, it is possible to reliably prevent the reflection from a vehicle other than the front vehicle, the reflection from the roadside object, etc. from being included in the reflection from the front vehicle as in the conventional case, which contributes to the improvement of the recognition accuracy of the front vehicle. be able to.

Further, the method for recognizing a forward vehicle according to the present invention includes a step of performing a motion estimation by a Kalman filter for the vehicle candidate point group judged to be a forward vehicle and utilizing it for the next matching processing by the matching unit. It has a feature.

According to such a configuration, the movement of the vehicle candidate point group determined to be the forward vehicle is predicted and used for the next matching process, so that more accurate vehicle recognition can be realized.

Further, in the method for recognizing a front vehicle according to the present invention, a predetermined gaze area is set for an image picked up by the image sensor, and X is calculated from the density of each pixel in the gaze area.
Direction edge histogram, a Y direction edge histogram orthogonal thereto, and a product of edge histograms in both directions are calculated to derive a local maximum point, thereby extracting the rectangular area as a vehicle candidate. It is characterized by that.

With such a configuration, when a rectangular area as a vehicle candidate is extracted from the edge histogram,
Other than the vehicle can be reliably excluded, and a rectangular area as a vehicle candidate can be accurately extracted.

[0025]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described with reference to FIGS. However, FIG. 1 is a block diagram, and FIGS. 2 to 10 are operation explanatory diagrams.

As shown in FIG. 1, a laser radar module 1 including a scan laser radar, its scanning mechanism and a light receiver, and a monocular CCD as an image sensor.
A CCD camera module 2 as an image processing unit including a camera is mounted in the vehicle, and the laser radar module 1 irradiates the vehicle in front while scanning the laser beam in the horizontal direction and receives the reflected light from the reflection point. The positions of a plurality of reflection points are specified by the camera module 2, the front of the vehicle is imaged by the camera module 2, and the reflection point data specified by the laser radar module 1 and the captured image obtained by the camera module 2 are acquired by the CPU 3. It is stored in the memory 4 including a RAM and the like. At this time, another memory may be provided to store the data from the laser radar module 1 and the data from the camera module 2 in different memories.

At this time, the relationship between the coordinate systems of the laser radar module 1 and the camera module 2 is set as shown in FIG. 2, for example. That is, as shown in FIG. 2, the camera coordinate system takes the camera optical axis as the Z axis with the lens center as the origin, and takes the X axis and the Y axis parallel to the x axis and the y axis on the image plane. Here, the focal length f is known.

In the laser radar coordinate system, the light receiver is the origin, the laser optical axis is the ZR axis, the YR axis is set in the vertical direction, and the XR axis is set in the horizontal direction. XR YR ZR X from coordinate system
The YZ coordinate system can be converted by a composite matrix ωT of the rotation vector ω and the translation vector T, and as described later,
In the present invention, the coordinates of the reflection point specified by the laser radar module 1 are converted into camera coordinates for verification. However,
The composite matrix ωT has been calibrated in advance,
It is known.

By the way, the laser radar module 1
Laser light is emitted in front of the vehicle, and the scanning mechanism scans the laser light horizontally at a predetermined angle (for example, 0.1 °), and the reflected light from the object in front of the vehicle is received by the light receiver. The distance between the irradiation of the laser light and the reception of the reflected light is detected from the distance between the vehicle and the preceding vehicle traveling in the same lane. The time required for such a single inter-vehicle distance detection process is about 100.
This is a short time of about ms, and this detection operation is repeated at regular intervals.

When a plurality of reflection points on the object are specified by the laser radar module 1 by such an operation, the vehicle widths of the vehicle widths are set to substantially equidistant positions based on the positions of the specified reflection points. The reflection point group existing within the spread range is set as the vehicle candidate point group. At this time, as described in Japanese Patent Laid-Open No. 6-309600,
The relative speed of the vehicle candidate point group with respect to the own vehicle is derived from the change in the distance between the own vehicle and the vehicle candidate point group, and when the derived relative speed is larger than a predetermined value, the own vehicle speed and the relative speed of the preceding vehicle candidate are Difference, the positional deviation amount of the preceding vehicle candidate with respect to the own vehicle in the vehicle width direction, etc. is used to calculate the certainty factor representing the certainty of being the preceding vehicle and the scanning direction of the laser beam. And, the calculation of the certainty factor is repeated every time the coordinates of the vehicle candidate point group and the relative speed in the coordinate system having the traveling direction of the own vehicle as the coordinate axis are calculated, and the difference between the certainty factor and the reference value is cumulatively added and the addition total is calculated. It may be determined whether or not the value is larger than another reference value, and when it is smaller, it may be determined that the vehicle candidate point group is traveling on the same lane, and when it is larger, it may be determined that the vehicle candidate point group is not on the same lane.

On the other hand, as shown in FIG. 3, the camera module 2 sets a rectangular vertical gaze region V-ROI that is long in the horizontal direction with respect to the picked-up image, and sets the density of the vertical gaze region V-ROI to the vertical direction. A vertical line segment is projected and extracted, and a set of local maximum points showing symmetry is selected from the extracted vertical line segments, and the set is used as both end candidates of the vehicle. Further, a horizontal gazing area H-ROI is set at the coordinate located at the center of the set of maximum points, and a horizontal line segment is extracted in the same manner, and if a large number are observed, the rectangular area is a vehicle candidate. Figure 4
Shows a time-series image of projected edge density values in the vertical gaze area V-ROI, and curves of symmetrical density at substantially equal intervals correspond to both ends of the vehicle.

However, in this case, the vertical gaze area V-ROI
However, the horizontal gaze area H-ROI must be set by the number of vehicle candidates, and the number of combinations increases as the maximum points in the vertical gaze area V-ROI increase. Then, the following improvements have been added. That is, as shown in FIG. 5, if the size of the vertical gaze region V-ROI that is long in the horizontal direction is appropriately set, a large number of horizontal edge components can be included in the vertical gaze region V-ROI. By projecting the horizontal edge component in the gazing area V-ROI, the total amount of the horizontal component can be detected to speed up the arithmetic processing.

Then, by the camera module 2, FIG.
As shown in, the vertical gaze area V-ROI of a predetermined size is set for the captured image, and the edge histogram EHx in the X direction (horizontal direction) from the grayscale of pixels in the vertical gaze area V-ROI, Y direction orthogonal to this (vertical direction)
Edge histogram EHy and the edge histogram products EHxy in both directions are calculated.

Referring to FIG. 5, (i) in FIG. 5 is a view in which the reflection of the radar is perspectively transformed and displayed on the image for reference, and the height represents the reflection intensity, and (ii) is A gaze area V-ROI set for image processing, (iii)
Indicates a value (edge histogram EHy) obtained by adding (projecting) vertical edge components in the vertical direction within the gazing area V-ROI, and (iv) similarly shows horizontal edge components in the vertical direction within the gazing area V-ROI. The added value (edge histogram EHx) is shown, and the vertical edge image and the horizontal edge image of the gazing area V-ROI are shown in (ix) and (x) for reference.

Further, (v) in FIG. 5 is a product (edge histogram EHxy) of the results of (iii) and (iv), and
The value becomes higher at the positions where many horizontal edges and vertical edges intersect, such as at both ends of the vehicle. In addition, (vi) is a smoothed version of (v) in order to make it easier to find the maximum point.
iii) shows the position of the maximum point.

At this time, since the rear glass of the vehicle and the vicinity of the rear bumper include a large amount of Y-direction edge components, the edge histogram EHy in the Y direction of the rear end surface of the vehicle has a large value, whereas the pole of a utility pole or a streetlight, For road structures such as guardrails, edge histogram E in the X direction
Hx increases, but edge histogram EH in the Y direction
y has a small value. Therefore, the edge histogram EHx in the X direction and the edge histogram EHy in the Y direction
By integrating and, the maximum points of the edge histogram other than the vehicle rear end surface can be suppressed to some extent, and conversely, the maximum points of the edge histogram of the vehicle rear end surface can be emphasized.

Therefore, the maximum point (xL, x) of the product EHxy of the edge histogram is preferentially calculated from the vehicle predicted position in the current frame calculated based on the vehicle recognized position in the previous frame.
R) is cut out as a vehicle rear end position candidate pair, and the edge histogram EV is obtained by horizontally counting Y-direction edges Ey equal to or larger than a predetermined threshold in the section (xL, xR).
y is created, and a vehicle upper and lower end candidate (yU, yD) is obtained from the degree of proximity of the vehicle upper (lower) end in consideration of the inter-vehicle distance and the estimated vehicle width at the maximum point in EVy. By doing so, it is possible to cope with the vehicle vertical position shift due to the change in the road gradient to some extent. The rectangle defined by these four points is a vehicle candidate. In this way, the rectangular area as the vehicle candidate is extracted.

Then, the CPU 3 coordinates-converts the vehicle candidate point group by the laser radar module 1 into the camera coordinate system of the camera module 2 and collates it with the rectangular area, and the vehicle candidate point group after the coordinate conversion substantially coincides with the rectangular area. When it does, it is determined that the vehicle candidate point group is a forward vehicle, and when they do not match, it is determined that the vehicle candidate point group is not a forward vehicle. The matching process by the CPU 3 corresponds to the matching unit, the determination process by the CPU 3 corresponds to the integrating unit, and these processes by the CPU 3 correspond to the recognizing unit in the present invention.

The collating process of the CPU 3 is conceptually represented as shown in FIG.
In the case where there are reflection points of a vehicle and a non-vehicle in the window set in the camera coordinate system, if the total reflection point is determined to correspond to, for example, the i-th vehicle, as shown in FIG. , The reflection points are labeled with the regular label i, and if there are no vehicles on the image, those reflection points are deleted. Further, when there are a plurality of regular labels existing in the window set in the image coordinate system, they are separated as different vehicles even in the same search window as shown in FIG. is there.

In this way, the system for integrating vehicle recognition by the laser radar and the camera as shown in FIG. 7 eliminates reflections from vehicles other than the vehicle in front, roadside objects, etc., and enables highly accurate recognition. Become. At this time, as a function of the CPU 3 corresponding to the integration unit, a Kalman filter that manages the position information of the front vehicle and other roadside structures and uses the processing result of the laser radar module 1 or the camera module 2 as an observation value is used. The position coordinates are estimated and used for the next matching process.

For example, as shown in FIG. 8 (a), (i) is an input image at a certain time t = t1, and (ii) is a recognition result by the image showing the positions of both ends of the vehicle. Is the x coordinate and the vertical axis is the time. (iii) shows the reflection points of the radar projected on the image coordinate system and expressed as a time series image.
v) is the result of labeling by comparing the recognition result of the image with the reflection point of the radar by the CPU 3, and FIG. 7B shows the time t = t0 shown in FIG. t = t1
It is a typical reflection point map observed in.

From FIG. 8, only the forward vehicle is brightly labeled and separated from other vehicles, and the reflection from the traveling vehicle or the parked vehicle near the forward vehicle, which is the first problem described in the prior art, is forward. It can be seen that it is possible to prevent the processing from being included in the reflection from the vehicle.

Next, FIG. 9 shows a scene in which a forward vehicle passes near a roadside object with many reflections. As shown in FIG. 9A, (i) is an input image at a certain time t = t3, and ii) ~ (iv)
Is the same as (ii) to (iv) of FIG. 8A, and FIG. 8B shows time t = t2 and t = t shown in FIG. 8A of (iii).
It is a typical reflection point map observed in 3. And
In the case of the laser radar alone, the recognition result is often pulled to the nearby roadside object, but as shown in FIG. 9, the roadside object can be removed by collation with the image. It can be seen that it is possible to prevent the processing from including the reflection from the roadside object, which is the second problem described above, in the reflection from the vehicle in front.

Further, FIG. 10 shows a recognition result for a situation where there is an overtaking from the radar detection range. At time t = t4, only the image recognizes an overtaking vehicle from the adjacent lane, and when the radar is alone, t = t5. Finally, both ends of the overtaking vehicle will be detected.

Since the reflection of the radar from the vehicle is partial in this section, it can be used as the distance information of this vehicle if the result of the collation matches the vehicle recognized in the image, and therefore, it can be used as the distance information of the radar alone. It is understood that the passing vehicle coordinates can be recognized at an early stage, and from FIG. 10, the interrupting vehicle at a short distance, which is the third problem described in the conventional technique, can be detected at an early stage.

As described above, according to the above-described embodiment,
The position of each reflection point specified by the laser radar module 1 and the rectangular area by the camera module 2 are the CPU
3 and the reflection point at the position corresponding to the rectangular area is determined to be the reflection point on the vehicle in front. Therefore, the characteristics of the laser radar that excels in distance measurement and the characteristics of the CCD camera that can accurately identify the vehicle and the other areas are used. However, by combining the vehicle recognition by the CCD camera with the vehicle recognition by the laser radar, it is possible to prevent the reflection from the vehicle other than the front vehicle or the roadside object from being included in the reflection from the front vehicle as in the conventional case. The vehicle in front can be recognized with high accuracy.

In the above embodiment, the case where the monocular CCD camera is used as the image sensor has been described, but it goes without saying that the image sensor is not limited to the monocular CCD camera described above.

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

[0049]

As described above, according to the first and fifth aspects of the invention, the laser radar is excellent in distance measurement, and the image sensor can accurately distinguish between the vehicle and the other parts. By integrating the above with the vehicle recognition by the laser radar, it is possible to prevent the reflection from the vehicle other than the front vehicle or the roadside object from being included in the reflection from the front vehicle as in the conventional case, and the front vehicle is highly accurate. It becomes possible to recognize, and it becomes possible to provide a more advanced and more comfortable driving support system.

Further, according to the inventions described in claims 2 and 6, the reflection points existing at substantially equidistant positions within the widening range of the vehicle width based on the positions of the respective reflection points specified by the laser radar. The group is defined as a vehicle candidate point group, and this vehicle candidate point group is coordinate-converted into the coordinate system of the image sensor and collated with a rectangular area. When the coordinate-converted vehicle candidate point group substantially matches the rectangular area, the vehicle Since the candidate point group is determined to be the front vehicle, it is possible to reliably prevent the reflection from the vehicle other than the front vehicle, the reflection from the roadside object, etc. from being included in the reflection from the front vehicle as in the conventional case. It becomes possible to contribute to the improvement of recognition accuracy.

According to the third and seventh aspects of the present invention, the movement of the vehicle candidate point group judged to be a forward vehicle is predicted and used in the next matching process, so that more accurate vehicle recognition is realized. It becomes possible to do.

Further, according to the present invention as defined in claims 4 and 8, when the rectangular area as the vehicle candidate is extracted from the edge histogram, the areas other than the vehicle can be reliably excluded, and the rectangular area as the vehicle candidate can be surely excluded. Can be accurately extracted.

[Brief description of drawings]

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

FIG. 2 is an operation explanatory diagram of the embodiment of the present invention.

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

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

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

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

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

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

FIG. 9 is an operation explanatory diagram of the embodiment of the present invention.

FIG. 10 is an operation explanatory diagram of the embodiment of the present invention.

FIG. 11 is an operation explanatory diagram of vehicle recognition by the scan laser radar which is the background of the present invention.

[Explanation of symbols]

1 Laser radar module 2 CCD camera module (image processing unit) 3 CPUs (collation unit, integration unit, recognition unit)

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) G01B 11/00 G08G 1/16 C G08G 1/16 H04N 7/18 J H04N 7/18 K G01S 17/88 AF term (reference) 2F065 AA01 AA12 BB05 BB15 CC11 FF04 JJ03 JJ26 QQ00 QQ24 QQ27 QQ29 QQ33 QQ36 QQ38 QQ43 QQ45 QQ51 5C054 AA01 AA04 CC02 FC12 HA04 5A180AA CC27

Claims (8)

[Claims] 1. A front vehicle recognizing device for recognizing a vehicle traveling in front of the host vehicle, wherein a plurality of laser beams are emitted in front of the host vehicle while being horizontally scanned and reflected light is received from a reflection point. A laser radar that identifies the position of the reflection point, an image sensor that images the front of the vehicle, an image processing unit that extracts a predetermined rectangular area including the rear end face image of the vehicle ahead from the image captured by the image sensor, The position of each of the reflection points specified by the laser radar is compared with the rectangular area by the image processing unit, and the reflection point at a position corresponding to the rectangular area is determined as a reflection point on the front vehicle, and the front vehicle is detected. And a recognition unit for recognizing the vehicle. 2. The vehicle candidate point group is configured such that the recognizing unit determines the reflection point groups existing at substantially equidistant positions based on the positions of the respective reflection points specified by the laser radar and within a wide range of the vehicle width. And a collating unit that transforms the vehicle candidate point group into the coordinate system of the image sensor and collates it with the rectangular area, and the vehicle when the vehicle candidate point group after coordinate conversion substantially matches the rectangular area. The front vehicle recognition device according to claim 1, further comprising: an integration unit that determines the candidate point group as a front vehicle. 3. The combination unit according to claim 2, wherein a motion prediction is performed by the Kalman filter for the vehicle candidate point group determined to be a forward vehicle, and the motion estimation is used for the next matching process by the matching unit. Recognition device for the vehicle ahead. 4. The image processing unit sets a predetermined gaze area for an image picked up by the image sensor, and an edge histogram in the X direction from the grayscale of each pixel in the gaze area, in the Y direction orthogonal thereto. 4. The rectangular area as a vehicle candidate is extracted by calculating a product of the edge histogram of No. 1 and the edge histogram in both directions and deriving a maximum point thereof. Recognition device for the vehicle ahead. 5. A front vehicle recognizing method for recognizing a vehicle traveling ahead of the own vehicle, wherein a laser radar irradiates the front of the own vehicle with laser light while scanning the same in a horizontal direction and reflects light from a reflection point. A step of receiving light and specifying the positions of the plurality of reflection points; and a step of extracting a predetermined rectangular area including the rear end face image of the front vehicle from the captured image obtained by capturing the front of the vehicle by the image sensor, A step of comparing the position of each reflection point specified by the laser radar with the rectangular area, and recognizing the front vehicle by determining that the reflection point at the position corresponding to the rectangular area is a reflection point on the front vehicle. The method for recognizing a forward vehicle, comprising: 6. The vehicle candidate point group is defined as the vehicle candidate point group, and the reflection point groups existing at substantially equidistant positions within a substantially widened range of the vehicle width based on the positions of the respective reflection points specified by the laser radar are set as vehicle candidate points. Coordinate conversion of the point cloud into the coordinate system of the image sensor and matching with the rectangular area; and when the vehicle candidate point cloud after coordinate conversion substantially matches the rectangular area, the vehicle candidate point cloud The method for recognizing a forward vehicle according to claim 5, further comprising: 7. The front vehicle according to claim 6, further comprising a step of performing a motion prediction by a Kalman filter for the vehicle candidate point group determined to be a front vehicle, and utilizing the motion estimation in a next matching process by the matching unit. Recognition method. 8. A predetermined gaze area is set for an image picked up by the image sensor, an edge histogram in the X direction from the grayscale of each pixel in the gaze area, an edge histogram in a Y direction orthogonal to the gray area, and both directions. 8. The vehicle ahead according to any one of claims 5 to 7, further comprising a step of calculating a product of an edge histogram of the vehicle and deriving a maximum point thereof to extract the rectangular area as a vehicle candidate. Recognition method.