(19) hp... mg. 035“ (JP)
`
`‘I2 Laid-Open Patent Application Publication (A)
`
`ii 1.-u|—0nu P-I Awlxaun NI-hr
`
`F21M 3/18
`B600 1/I-1
`FZIM 3/os
`
`Ho4N 7/I8
`
`_
`
`J
`
`Toklcai-H 6-267303
`
`(43)Dne ofhblxmon Sqnunball. I994
`
`Iechnbgy Nome hdiutun
`
`Reqnestfasltbsraamrfixannmuom Nukupestcd No.o!'Ch:ms 101. (Total. l-tpogs)
`
`(21) Patent appnaummnha rrs.57o3o
`
`(22) Filmgdate March 17. 1993
`
`(7l)AwI<- 000003207
`Toyota Motor Corporation
`-" T°’°" """ T°"“""‘ ‘“°"' k‘’'
`(72) lmunor Masasln MIZUKOSHI
`Toyota Motor Corporation
`1. Toyota-tho. Toyota-shL
`
`(74)Agent
`
`Patent
`
`IMNAKASHIMA (and 2 othus)
`
`(54) [l'It|e ot the invention] Headlamp Device for Vehicle
`
`(57)[Abstract]
`[Object] To reliably prevent the creation of glare on other
`vehicles.
`[Structure] A TV camera 22 that images the conditions
`ahead of a vehicle and a radar 80 that measures inter-
`vehicle distance to another vehicle is mounted in a vehicle
`10. The radar 80 is rotated in the direction of arrow A or in
`the direction of arrow B by actuators (not pictured). Based
`on the image signals output from the TV camera 22,
`another vehicle is recognized by an image processing
`device (not pictured), and the position of each vehicle in
`the image is detennined. Next, based on the position of
`each vehicle that has been detennined, the direction in
`which each vehicle is present is found, the radar 80 is
`rotated so that the other vehicle is contained within the
`radar 80 detection area, and the inter-vehicle distance to
`each vehicle is measured. Additionally, based on the
`measured inter—vehicle distances, a light-shielding cam
`(not shown) provided on a headlamp 18, 20 is rotated,
`thereby controlling the illumination range at the lamps so
`as not to create glare on other vehicles.
`
`Mercedes—Benz USA, LLC, Petitioner — Ex. 1008
`
`1
`
`Mercedes-Benz USA, LLC, Petitioner - Ex. 1008
`
`

`
`[Claims]
`[Claim 1]
`A headlamp device for a vehicle having:
` a headlamp, at least one of illumination direction and
`illumination range thereof being changeable;
` an imaging means that images the conditions ahead of a
`vehicle and outputs image signals;
` a detecting means that, based on image signals output from
`said imaging means, detects the direction in which another
`vehicle is present;
` a measuring means that, based on the direction of another
`vehicle detected by said detecting means, measures an inter-
`vehicle distance to said another vehicle; and,
` a controlling means that, based on at least the inter-vehicle
`distance to said another vehicle measured by said measuring
`means, controls at least one of illumination direction and
`illumination range of said headlamp so as not to create glare
`on said another vehicle.
`[Detailed description of the invention]
`[0001]
`[Field of industrial use] The present invention relates to a
`headlamp device for a vehicle, and in particular, to a
`headlamp device for a vehicle that, while a vehicle is traveling,
`controls the light distribution of a headlamp that illuminates
`the area ahead of a vehicle.
`[0002]
`[Prior art] Headlamps are placed as a pair on the right side
`and left side of the front end of a vehicle, and light up in
`cases in which it is difficult to view the conditions ahead such
`as at night, thereby enhancing a driver’s forward visibility.
`These headlamps generally are configured so as to be
`switchable between only two stages of illumination range,
`high beam and low beam, and in cases in which other
`vehicles such as leading vehicles or oncoming vehicles are
`present, low beam is often selected so as not to create a
`glare that is blinding and discomforting to drivers of other
`vehicles. Nonetheless, in cases in which the inter-vehicle
`distance to a leading vehicle is long, for example, there is the
`problem that with low beam, a driver continuously sees dark
`portions outside the illumination range of the headlamps, and
`with high beam, glare is created on leading vehicles and the
`like, thereby making it difficult to illuminate the area ahead
`within an appropriate range constantly.
`[0003] Therefore, it has been proposed that a visor for
`shielding illuminating light is provided inside a headlamp, said
`visor being moved so as not to create glare on other vehicles
`and also so as a sufficient illumination range can be obtained,
`thereby controlling the position of the boundary between an
`illuminated area and a not-yet-illuminated area (hereinafter,
`this boundary is called a cutoff line). Further, it has been
`proposed, as a technology to control the position of the cutoff
`line so as not to create glare on other vehicles, that the
`conditions ahead of a vehicle is imaged by a CCD camera or
`the like; a leading vehicle is recognized and the inter-vehicle
`distance to the leading vehicle is detected based on image
`signals output from the CCD camera; and the headlamp light
`distribution
`is controlled according
`to
`the
`inter-vehicle
`distance (See Japanese Unexamined Patent Application
`Publication S62-131837).
`[0004] Detection of inter-vehicle distance based on images
`representing the conditions ahead of a vehicle is performed
`based on the principle that as the inter-vehicle distance to
`another vehicle increases, the position of the other vehicle in
`said image moves toward the top of the image, and the
`farther toward the top of the image the other vehicle is
`positioned, the greater the inter-vehicle distance is judged to
`
`be. Further, since it is difficult to detect other vehicles
`themselves in conditions such as at night, if a leading vehicle,
`tail lamps, and if an oncoming vehicle, headlamps, are
`detected, and inter-vehicle distance is judged based on the
`height position of the lamps or the interval between the lamps
`in the image.
`[0005]
`[Problems the invention is to solve] Nonetheless, the height
`position of said tail lamps and headlamps relative to the road
`surface and the interval between the lamps vary by vehicle
`model. Therefore, in detecting inter-vehicle distance based on
`the height position of tail lamps or headlamps, or the interval
`between lamps, even if the actual inter-vehicle distance is the
`same, the inter-vehicle distance detection results can vary
`according to the vehicle model of the other vehicle. Further,
`the height position of lamps in an image will change in cases
`in which, for example, the relative height position relative to
`other vehicles changes due to the incline of the road surface
`or the like. Accordingly, inter-vehicle distance cannot be
`detected accurately by the lamp height position and lamp
`interval in an image, and thus there is a possibility of creating
`glare on other vehicles even if headlamp light distribution is
`controlled based on the detected inter-vehicle distance to
`other vehicles.
`[0006] The present invention was created with consideration
`of the aforementioned situations, and an object thereof is to
`obtain a headlamp device for a vehicle that can reliably
`prevent the creation of glare on another vehicle.
`[0007]
`[Means for solving the problems] In order to achieve the
`above object, the headlamp device for a vehicle has: a
`headlamp, at
`least one of
`illumination direction and
`illumination range thereof being changeable; an imaging
`means that images the conditions ahead of a vehicle and
`outputs image signals; a detection means that, based on
`image signals output from said imaging means, detects the
`direction in which another vehicle is present; a measuring
`means that, based on the direction of another vehicle
`detected by said detecting means, measures an inter-vehicle
`distance to said another vehicle; and a controlling means that,
`based on at least the inter-vehicle distance to said another
`vehicle measured by said measuring means, controls at least
`one of illumination direction and illumination range of said
`headlamp so as not to create glare on said another vehicle.
`[0008]
`[Function] In the present invention, the direction in which
`another vehicle is present is detected based on image signals
`obtained by imaging conditions ahead of a vehicle, the inter-
`vehicle distance to another vehicle based on the detected
`direction of another vehicle is measured by a measuring
`means, and at least based on the inter-vehicle distance to
`another vehicle, at least of one of illumination direction and
`illumination range of the headlamps is controlled so as not to
`create glare on another vehicle. Commonly known millimeter
`wave radar, laser radar and the like, for example, can be
`applied as said measuring means. Further, it is also possible
`to apply,
`for example, a geodimeter, which measures
`distance using light interference and is used in surveying and
`the like; or a Tellurometer, which measures the round-trip
`time for microwaves reflected from a measurement target and
`finds distance based on phase comparison.
`
`2
`
`

`
`(3)
`
`
`
`Tokkai H6-267303
`
`image processing device 48 and outputs the vehicle speed V
`detection results.
`[0014] Further, a radar 80 is disposed as a measuring means
`on the inside of the front grill of the vehicle 10. In the present
`embodiment, a millimeter radar with a detection range width
`approximately the size of one lane travelled by a vehicle is
`used as the radar 80. An actuator 82 (see Figure 4) that
`rotates the radar 80 in the direction of arrow A and the
`direction of arrow B in Figure 1 is coupled to the radar 80.
`Rotation of radar 80 by the actuator 82 in the direction of
`arrow A or the direction of arrow B in Figure 1 contains
`another vehicle present in any direction ahead of the vehicle
`10 within the detection area and allows detection of the inter-
`vehicle distance to the other vehicle. The radar 80 is
`connected to an input port 58 of a control device 50 (see
`Figure 4) and outputs the result of inter-vehicle distance
`detection to the control device 50. Further, said actuator 82 is
`connected to an output port 64 of the control device 50 and
`rotates the radar 80 in the direction of arrow A or the direction
`of arrow B by just the rotation angle instructed by the control
`device 50.
`[0015] As shown in Figure 2 and Figure 3, the headlamp 18 is
`a projector-type headlamp equipped with a convex lens 30, a
`bulb 32, and a lamp housing 34. The lamp housing 34 is fixed
`generally horizontally to the frame (not pictured) of the vehicle
`10, the convex lens 30 is fixed to one of the lamp housing 34
`openings, and at another opening bulb 32 is affixed via a
`socket 36 such that its illumination spot is positioned on the
`optical axis L of the convex lens 30 (the central axis of the
`convex lens 30).
`[0016] A reflector 38 with an elliptical reflective surface is
`formed on the bulb side of the interior of the lamp housing 34,
`and light emitted from the bulb 38 [sic] is reflected by the
`reflector 38 and focused between the convex lens 30 and the
`bulb 32. Actuators 40, 42 are placed in proximity to the focal
`point. An actuator 40 is equipped with a light-shielding cam
`40A with the axle thereof supported rotatably around a
`rotational axis 44 fixed inside the lamp housing 34 along the
`width of the vehicle, and a gear 40B is secured to the light-
`shielding cam 40A. A gear 40C secured to the drive axle of a
`motor 40D engages the aforementioned gear 40B. The motor
`40D is connected to a driver 64 of the control device 50.
`[0017] Further,
`the actuator 42,
`like
`the actuator 40,
`comprises a light-shielding cam 42A with the axle thereof
`supported rotatably around said rotational axis 44, a gear 40B
`secured to the light-shielding cam 40A [sic], a motor 42D, and
`a gear 40C [sic] which is secured to the drive axle of the
`motor 42D and engages the aforementioned gear 40B [sic].
`The motor 40D [sic] is also connected to the driver 64 of the
`control device 50. The light of the bulb 32 reflected and
`
`[0009] The aforementioned measuring means, because the
`detection value of inter-vehicle distance does not vary due to
`the vehicle model of the vehicle that is the detection target or
`due to changes in the relative height relative to the vehicle
`that is the detection target, can accurately measure inter-
`vehicle distance by detecting the lamp height position or lamp
`interval in an image and comparing this to cases in which the
`inter-vehicle distance to another vehicle is judged or the like,
`and based on this accurately measured inter-vehicle distance,
`by controlling at least one of illumination direction and
`illumination range of the headlamps, can prevent creation of
`glare on another vehicle.
`[0010] Moreover, in cases in which radar is used as a
`measuring means to detect inter-vehicle distance to another
`vehicle, said inter-vehicle distance to another vehicle can be
`measured by using a radar with a relatively small output and
`sharp directivity, for example, and by orienting the radar so
`that another vehicle whose position has been detected falls
`within the detection range of the radar. Further, it is also
`acceptable to obtain the inter-vehicle distance of a specific
`other vehicle based on the position of the other vehicle after
`the distance relative to all objects present in said detection
`range has been detected using a radar with a relatively large
`output and broad detection range. Nonetheless, because it is
`desirable for a measuring means to be installed in a vehicle
`to be compact and inexpensive, and because its output
`cannot be very large, it is preferable to use a radar with a
`relatively small output and sharp directivity to detect inter-
`vehicle distance, as discussed previously.
`[0011]
`[Embodiments]
`Hereinafter, an embodiment of the present invention is
`described in detail with reference to drawings. As shown in
`Figure 1, an engine hood 12 is disposed on the top surface of
`the front body 10A of a vehicle 10, and a front bumper 16 is
`affixed to the front end of the front body 10A across the width
`of the vehicle from one end to the other end. A pair of
`headlamps 18, 20 is disposed at each end of the width of the
`vehicle, between the front bumper 16 and the front edge of
`the engine hood 12.
`[0012]
`A windshield glass 14 is provided near the back edge of the
`engine hood 12, and a room mirror 15 is provided in proximity
`to the portion corresponding to the top of the windshield glass
`14 on the inside of the vehicle 10. A TV camera 22 for
`imaging the conditions ahead of a vehicle is disposed in
`proximity to the room mirror 15. The TV camera 22 is
`connected to an image processing device 48 (see Figure 4).
`In the present embodiment, a TV camera equipped with a
`CCD element that simply detects only the amount of light and
`outputs image signals representing black-and-white images is
`used as the TV camera 22.
`[0013] The TV camera 22 is preferably disposed in a position
`as close as possible to the driver perspective (the "eye point")
`so that the configuration of the road ahead of the vehicle can
`be accurately recognized and the visual sensation of the
`driver can be more closely matched. Further,
`the
`configuration of the road in the present embodiment includes
`the configuration of the path of travel, for example, a road
`configuration corresponding to a one-vehicle lane formed by
`a center line or a curb, etc. Further, a speedometer (not
`pictured) is placed in the vehicle 10, and a vehicle speed
`sensor 66 (see Figure 4) that detects the vehicle speed V of
`the vehicle 10 is attached to the cable of the speedometer not
`pictured. The vehicle speed sensor 66 is connected to the
`
`3
`
`

`
`(4)
`
`
`
`Tokkai H6-267303
`
`processing is described with reference to the flow chart in
`Figure 6.
`[0025] An example image (image 120) imaged by the TV
`camera 22 while the vehicle 10 is traveling on a road 122 and
`that generally matches the image visible to the driver is
`shown in Figure 8(A). The road 122 is provided with a white
`line 124 on both sides of the lane in which the vehicle 10 is
`traveling. The position of each pixel in the above-mentioned
`image is specified by coordinate system coordinates (Xn, Yn)
`determined by the intersection of the X-axis and Y-axis set on
`the image. Next, recognition of another vehicle, including a
`leading vehicle, is performed based on this image.
`[0026] In Step 300 of the flow chart of Figure 7, an area
`on the image as shown in
`having a predetermined width
`Figure 9 is set as a white line detection window area
`sd. In
`the present embodiment, when the vehicle 10 is traveling at
`night, in consideration of the fact that an image can only be
`detected up to about 40-50m ahead of the vehicle 10, white
`line detection is not performed at distances exceeding 60m
`ahead of the vehicle 10. Further, the probability that a leading
`vehicle will be present in the lower area of an image is low.
`sd, so as to
`Therefore, the white line detection window area
`make detection up to 60m ahead of the vehicle 10 possible, is
`sd with the area
`set as a white line detection window area
`above a predetermined horizontal line 140 and the area
`below a lower limit line 130 removed.
`sd is
`[0027] Next, in Step 302, content in the window area
`differentiated for brightness, and the peak point (maximum
`point) of these derivatives is extracted as an edge point that
`is a white line candidate point. That is, in the window area
`sd, brightness is differentiated for each pixel in a vertical
`direction (direction of arrow A in Figure 9) and a horizontal
`direction from the lowest position pixel to the highest position
`pixel, and the peak points of derivatives with a large variation
`in brightness are extracted as edge points. Thus, as an
`example, continuous edge points are extracted as shown by
`sd of Figure 9.
`the dotted line within the window area
`[0028] In Step 304, straight line approximation processing is
`performed. This processing approximates straight lines by
`using Hough conversion to convert the edge points extracted
`in white
`line candidate extraction processing, and
`approximate straight lines 142, 144 along the line assumed to
`be a white line are found. Next, Step 305 finds the point of
`intersection PN (where X coordinate value = XN) for the
`approximate straight lines thus determined, as well as the
`horizontal displacement A (A
`XN - X0) of the point of
`intersection PN thus determined from a point of intersection
`PO (where X coordinate value = XO) for an approximate
`straight line in the case of
`
`Ｗ
`Ｗ
`
`focused by the reflector 38 is shielded by the light-shielding
`cams 40A, 42A of the actuators 40, 42, and all other light is
`projected from the convex lens 30.
`[0018] Said light-shielding cams 40A, 42A have a cam shape
`wherein the distance from the rotational axis 44 to the outer
`boundary changes continuously along the circumferential
`direction, and are rotated individually by the motors 40D, 42D
`operated according to signals from the control device 50. The
`position of the boundary where the light of the bulb 32 is
`divided into transmitted light and shielded light is changed
`vertically according to the rotation of the light-shielding cams
`40A, 42A. This boundary appears as the cutoff line that is the
`boundary between light and dark of the light distribution
`ahead the vehicle 10.
`[0019] As shown in Figure 22, said boundary formed by the
`light-shielding cam 40A appears as a cutoff line 70 on the
`right side of the vehicle width within the area illuminated by
`the headlamp 18, and by rotating the light-shielding cam 40A,
`the position of the cutoff line 70 moves in parallel from the
`position corresponding to the highest position (in Figure 22,
`the position shown by a solid line as the cutoff line 70, a
`position at or lower than the so-called high beam) to the
`position corresponding to the lowest position (in Figure 22,
`the position shown by an imaginary line, a position along the
`so-called low beam).
`[0020] Further, said boundary formed by the light-shielding
`cam 42A appears as the cutoff line 72 on the left side of the
`vehicle width within the illuminated area, and by rotating the
`light-shielding cam 42A, the position of the cutoff line 72
`moves in parallel from the highest position (in Figure 22, the
`position shown by a solid line as the cutoff line 72, a position
`at or lower than the so-called high beam) to the lowest
`position (in Figure 22, the position shown by an imaginary line,
`a position along the so-called low beam).
`[0021] Further, because the headlamp 20 has the same
`configuration as the headlamp 18, a detailed description is
`omitted, but, as shown in Figure 4, actuators 41, 43 are
`attached, and the position of the cutoff line on the left side
`and the position of the cutoff line on right side of the
`illuminated area
`respectively
`is moved
`independently
`according to operation of the actuators 41, 43.
`[0022] As shown in Figure 4, the control device 50 is
`configured by containing a read-only memory (ROM) 52; a
`random access memory (RAM) 54; a central processing
`device (CPU) 56; an input port 58; an output port 60; and a
`data bus, control bus, or other such bus 62 that connects the
`aforementioned components. Moreover, a map and a control
`program, discussed below, are stored in said ROM 52.
`image
`[0023] The vehicle speed sensor 66 and
`the
`processing device 48 are connected to the input port 58. As
`discussed below, the image processing device 48 performs
`image processing of images made by the TV camera 22
`based on signals input from the TV camera 22 and the control
`device 50. The output port 60 is connected via the driver 64
`to the actuators 40, 42 of the headlamp 18 and the actuators
`41, 43 of the headlamp 20. Further, the output port 60 is also
`connected to the image processing device 48.
`[0024] Next, the function of the present embodiment is
`described with reference to the flow charts of Figure 5
`through Figure 7. When a driver turns on the light switch (not
`pictured) in the vehicle 10, thereby lighting the headlamps 18,
`20, the main control routine shown in Figure 5 is implemented
`at predetermined time intervals. In Step 200 of the main
`control routine, leading vehicle recognition processing is
`performed, and a leading vehicle driving ahead of the host
`vehicle 10 is recognized. This leading vehicle recognition
`
`γ
`
`＝
`
`Ｗ
`
`Ｗ
`
`Ｗ
`
`Ｗ
`
`4
`
`

`
`(5)
`
`
`
`Tokkai H6-267303
`
`α
`
`the correction width
`displacement amount A),
`is
`R
`decreased by decreasing the right side gain GR, and the
`correction width
`L is increased by increasing the left side
`gain GL. This change in correction width is shown as an
`image in Figure 15.
`[0034] In Step 324, the area enclosed by the approximate
`straight lines 142, 144 with their positions corrected by the
`L,
`R as determined is set as the leading
`correction widths
`vehicle recognition area WP.
`[0035] On the other hand, in cases in which the judgment of
`Step 314 is positive, it is judged that the road is a left-curving
`road, transition is made to Step 326, and the vehicle speed V
`of the vehicle 10 is read. In Step 328, the left and right
`L',
`R' that correspond to the read-out
`correction values
`vehicle speed V are determined using the map shown in
`Figure 12, and in Step 330, the gains GL, GR corresponding
`to the displacement amount A are determined. That is,
`because the likelihood that a leading vehicle is present on the
`left side is high when the road is a left-curving road with a
`small radius of curvature (large displacement amount A), the
`R is decreased by decreasing the right
`correction width
`side gain GR using the map shown in Figure 16, and the
`correction width
`L is increased by increasing the left side
`gain GL using the map shown in Figure 17.
`[0036] Next, in Step 332, final left and right correction widths
`R,
`L of the window area are determined based on the
`correction values
`R',
`L' and gains GL, GR as determined,
`and in Step 334, the leading vehicle recognition area WP is
`set as the area enclosed by the approximate straight lines
`142, 144 with their positions corrected by the left and right
`correction widths
`R,
`L as determined. When the leading
`vehicle recognition area WP has been set as described
`above, transition is made to Step 336.
`[0037] In Step 336, horizontal edge detection processing
`within the leading vehicle recognition area WP is performed
`as leading vehicle recognition processing. In horizontal edge
`detection processing, first, as with edge detection processing
`in Step 302, detection of the horizontal edge points is
`performed within the vehicle recognition area WP. Next, the
`detected horizontal edge points are differentiated in a
`P with a derivative
`horizontal direction, and the peak point
`that exceeds a predetermined value is detected (see Figure
`8(B)). This horizontal edge is highly likely to appear when a
`leading vehicle is present.
`[0038] Next, in Step 338, the positional coordinates of a
`leading vehicle is computed. First, vertical edge detection
`processing is performed. When there are multiple peak points
`P of derivatives of the horizontal edge points, the window
`areas WR, WL for detecting vertical lines are set in order from
`the peak point
`P positioned at the bottom portion of the
`image so as to include both ends of the horizontal edge
`points included in the peak points
`P (see Figure 8(C)). The
`vertical edge is detected within these window areas WR, WL,
`and in cases in which the vertical lines 138R, 138L are stable
`and are detected, it is judged that a leading vehicle is present
`in the area sandwiched between the window areas WR, WL.
`
`α
`α
`α
`α
`α
`α
`α
`
`α
`α
`α
`α
`
`Ｅ
`
`α
`
`α
`
`Ｅ
`
`Ｅ
`
`Ｅ
`
`a straight road, determined in advance as a reference. This
`displacement amount A corresponds to the degree of curve of
`the road 122.
`[0029] Next, in Step 306, it is judged whether or not the road
`122 is a generally straight road by judging whether or not the
`A
`A1.
`displacement amount A is within the range of A2
`The judgment reference value A1 is a reference value
`representing the boundary between a straight road and a
`right-curving road, and the judgment reference value A2 is a
`reference value representing the boundary between a straight
`road and a left-curving road. In cases in which a straight road
`is judged in Step 306, the vehicle speed of the host vehicle
`10 is read out in Step 308.
`R that
`L,
`[0030] Next, in Step 310, correction widths
`correct the position of the approximate straight lines are
`determined by setting a leading vehicle recognition area WP
`that recognizes leading vehicles corresponding to the read-
`out vehicle speed V. Because the radius of curvature of a
`road on which a vehicle can turn when traveling at high
`speeds is large, travel on a generally straight road is an
`allowable approximation, but because the radius of curvature
`when traveling at low speeds is small, even if the road directly
`ahead of the vehicle is close to a generally straight road, in
`cases in which the radius of curvature of the road far ahead is
`decreasing, it is possible that the vehicle may deviate from
`the leading vehicle recognition area WP. Therefore, using a
`map such as the one shown in Figure 12, the values of said
`L,
`R are determined so as to increase
`correction widths
`as vehicle speed V decreases.
`[0031] Next, in Step 312, the area enclosed by the lower limit
`line 130 and the approximate straight lines 142, 144 with their
`L,
`R is set as
`positions corrected by corrections widths
`the leading vehicle recognition area WP (see Figure 10).
`Moreover, for this leading vehicle recognition area WP as well,
`as travel speed decreases and correction widths
`L,
`R
`change in response to changes in vehicle speed V, the area
`is made larger (see Figure 11).
`[0032] On the other hand, when judgment is negative in Step
`A2, it is
`306, in Step 314, by judging whether or not A
`judged whether the road is a right-curving road or a left-
`curving road. In cases in which judgment is positive, the road
`is judged to be a right-curving road; the vehicle speed V of
`the vehicle 10 is read out in Step 316; and in Step 318, the
`correction values
`L',
`R' for the correction widths
`L,
`R
`that correspond
`to
`the read-out vehicle speed V are
`determined using the map shown in Figure 12. Next, in Step
`320, the maps in Figure 13 and Figure 14 are used to
`determine gains GL, GR for determining corrections widths
`R,
`L for left and right approximate straight lines that
`correspond to the displacement amount A that represents the
`degree of curvature. In Step 322, final left and right correction
`widths for the window area are determined based on the
`correction values
`R',
`L' and gains GL, GR as determined.
`[0033] Because the road in this instance is a curved road, it is
`left-right asymmetrical, and the approximate straight lines 142,
`144 have different inclinations. Thus, the left and right
`correction widths
`R,
`L are set as independent values.
`That is, when the road is a right-curving road with a small
`radius of curvature (large displacement amount A), the
`likelihood that a leading vehicle is present on the right side is
`R is increased by
`high. Accordingly, the correction width
`increasing the right side gain GR (see Figure 13), and the
`correction width
`L is decreased by decreasing the left side
`gain GL (see Figure 14). Further, when the road is a right-
`curving road with a
`large radius of curvature (small
`
`≧
`
`≧
`α
`
`α
`
`α
`
`α
`
`α
`＞
`α
`
`α
`α
`α
`
`α
`
`α
`α
`α
`α
`
`α
`α
`α
`α
`α
`
`α
`
`5
`
`

`
`[0039] Next, vehicle width is found by finding the horizontal
`interval between the vertical lines 138R, 138L detected with
`the window areas WR, WL, respectively, and the coordinates
`of the vehicle width center are found as the coordinates of the
`vehicle position in the image. The leading vehicle recognition
`processing is concluded as described above, and transition is
`made to Step 202 of the main control routine shown in Figure
`5.
`[0040] In Step 202, oncoming vehicle recognition processing
`is performed. Oncoming vehicle recognition processing is
`described with reference to the flow chart of Figure 7. Step
`404 reads
`the horizontal displacement amount A, as
`determined in the foregoing leading vehicle recognition area
`N of an
`setting processing, between the intersection point
`approximate straight line and the intersection point
`0 of an
`approximate straight line for a reference straight road (see
`Step 305). Next, in Step 406, it is judged whether or not the
`displacement amount A is in the range of A2
`A
`A1, and in
`cases in which the judgment is positive, the road 122 is
`judged to be a generally straight road, the vehicle speed V of
`the vehicle 10 is read out in Step 408. Next, Step 410
`RO that corrects
`determines the right side correction width
`the position of the approximate straight line for setting the
`oncoming vehicle recognition area WPO corresponding to
`vehicle speed V as read. This correction width
`RO, as with
`the aforementioned
`R,
`L
`for
`the
`leading vehicle
`recognition area WP, is increased when driving at low speeds
`and decreased when driving at high speeds, using the map
`shown in Figure 18. Next, in Step 412, the oncoming vehicle
`recognition area WPO for recognizing an oncoming vehicle is
`determined using the lower limit line 130, the approximate
`straight line 144, and the correction width
`RO determined
`(see Figure 19).
`[0041]
`On the other hand, in cases in which the judgment of Step
`406 is negative, in Step 414, it is judged whether or not
`A2. In cases in which the
`displacement amount A is A
`judgment of Step 406 is positive, it is judged that the road is a
`right-curving road, the vehicle speed V of the vehicle 10 is
`read out in Step 416, and next in Step 418, the map in Figure
`RO' for the
`18 is used to determine the correction value
`correction width
`RO corresponding to the vehicle speed V
`as read. Next, in Step 420, the gain GRO for determining the
`RO is determined using the map of Figure
`correction width
`20, and Step 422 determines the correction width
`RO for
`correcting the position of the approximate straight line 144
`based on the correction value
`RO' and gain GRO as
`determined. In Step 424, the correction width
`RO as
`determined is used to determine the oncoming vehicle
`recognition area WPO for recognition processing of oncoming
`vehicles.
`[0042] On the other hand, in cases in which the judgment of
`Step 414 is negative, the road is judged to be a right-curving
`road, transition is made to Step 426, and the vehicle speed V
`of the vehicle 10 is read. Next, in Step 428, a correction value
`RO' is determined based on the vehicle speed V as read
`and the map of Figure 18, and in Step 430, gain GRO
`corresponding to the displacement amount A is det