JP2017100593A - Outside-car environment recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform proper light distribution control to a head lamp and avoid a recognition error of a visual guidance lamp.SOLUTION: An outside-car environment recognition device is configured so that a computer functions as: an oncoming vehicle extraction part for extracting head lamp candidates of oncoming vehicles from an image, with a condition in which, at least luminance of a pixel is equal to or more than a prescribed luminance threshold; a visual guidance lamp extraction part for extracting a visual guidance lamp from the extracted head lamp candidates, for excluding the visual guidance lamp from the head lamp candidates; an oncoming vehicle recognition part for recognizing oncoming vehicles based on the head lamp candidates; and a travel scene determination part for, based on extraction of the visual guidance lamp, determining whether or not, a scene is a scene in which the visual guidance lamp exists, and when it is determined that, the scene is the scene in which the visual guidance lamp exists, identifying an existence area 248 in which the visual guidance lamp exists. The oncoming vehicle extraction part sets the luminance threshold in the existence area to be higher than that in areas other than the existence area, when it is determined that the scene is the scene in which the visual guidance lamp exists.SELECTED DRAWING: Figure 13

Description

  The present invention relates to a vehicle exterior environment recognition device that performs light distribution control of a headlamp according to a vehicle exterior environment.

  Conventionally, a three-dimensional object such as a vehicle positioned in front of the host vehicle is detected, and a collision with a preceding vehicle is avoided (collision avoidance control), or the distance between the preceding vehicle and the preceding vehicle is controlled to be kept at a safe distance ( (Cruise control) technology is known (for example, Patent Document 1).

  In addition, for safe driving at night, the adoption of an autolight function that automatically turns on the headlamp when the brightness outside the vehicle is not sufficient is in progress. Also, a technique is known in which the brightness of the surrounding environment is determined based on exposure information obtained from a monitor sensor and used for autolight (for example, Patent Document 2).

Japanese Patent No. 3349060 JP-A-11-187390

  By adopting the autolight as described above, the headlamp can be turned on uniformly according to the environment outside the vehicle regardless of the judgment of the driver. However, even when the outside of the vehicle is dark, the oncoming vehicle must not be dazzled (glare) by the high beam. Therefore, when a high brightness area (headlamp) is extracted from the image acquired through the imaging device and the oncoming vehicle is recognized based on the high brightness area, the high beam is switched to the low beam (HBA), or the area where the oncoming vehicle is located. The high beam is irradiated (ADB) avoiding the above.

  However, if a high-intensity gaze guidance lamp (snow pole) located on the shoulder of the traveling path is mistakenly recognized as a headlamp, a high beam is emitted to the vicinity of the gaze guidance lamp that should originally be illuminated with a high beam. Control such as not working.

  The present invention has been made in view of such a problem, and an object of the present invention is to provide a vehicle environment recognition device that can avoid erroneous recognition of a line-of-sight guide light and can appropriately control light distribution of a headlamp.

  In order to solve the above-described problem, the vehicle exterior environment recognition apparatus according to the present invention is a vehicle on the other hand that extracts a headlamp candidate of an oncoming vehicle from an image on the condition that at least the luminance of a pixel is equal to or higher than a predetermined luminance threshold. An extraction unit, a line-of-sight guide light extraction unit that extracts a line-of-sight guide light from the extracted headlamp candidates and excludes the line-of-sight guide light from the headlamp candidates, and an oncoming vehicle recognition unit that recognizes an oncoming vehicle based on the headlamp candidates Based on the extraction of the line-of-sight guide light, it is determined whether the scene is a scene where the line-of-sight guide light is present. The oncoming vehicle extraction unit functions as a determination unit. And wherein the Kusuru.

  The oncoming vehicle extraction unit may lower the brightness threshold in the area other than the existing area if it is determined that the scene includes the line-of-sight guide light.

  According to the present invention, it is possible to avoid misrecognition of the line-of-sight guide light and appropriately control the light distribution of the headlamp.

It is the block diagram which showed the connection relation of the external environment recognition system. It is the functional block diagram which showed the schematic function of the external environment recognition apparatus. It is a flowchart which shows the flow of a vehicle exterior environment recognition process. It is explanatory drawing for demonstrating a luminance image and a distance image. It is explanatory drawing for demonstrating a detection range. It is explanatory drawing for demonstrating the brightness | luminance image from which exposure time differs. It is explanatory drawing for demonstrating the installation state of a gaze guidance lamp. It is explanatory drawing for demonstrating the characteristic of a color. It is explanatory drawing for demonstrating the characteristic of a color. It is explanatory drawing for demonstrating a distant oncoming vehicle. It is explanatory drawing which shows operation | movement of a gaze guidance light extraction part. It is explanatory drawing for demonstrating re-determination of a line-of-sight guide light. It is explanatory drawing for demonstrating an existing area. It is explanatory drawing for demonstrating the characteristic of a color. It is explanatory drawing for demonstrating operation | movement of a driving scene determination part. It is explanatory drawing which showed an example of the determination process of a driving scene determination part. It is explanatory drawing which showed an example of the determination process of a driving scene determination part. It is explanatory drawing which showed operation | movement of the light distribution control part. It is explanatory drawing explaining the light distribution control of ADB. It is an upper surface figure for demonstrating a cut line angle.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The dimensions, materials, and other specific numerical values shown in the embodiment are merely examples for facilitating understanding of the invention, and do not limit the present invention unless otherwise specified. In the present specification and drawings, elements having substantially the same function and configuration are denoted by the same reference numerals, and redundant description is omitted, and elements not directly related to the present invention are not illustrated. To do.

(External vehicle environment recognition system 100)
FIG. 1 is a block diagram showing a connection relationship of the external environment recognition system 100. The vehicle exterior environment recognition system 100 includes an imaging device 110, a vehicle exterior environment recognition device 120, and a vehicle control device (ECU: Engine Control Unit) 130.

  The imaging device 110 is configured to include an imaging element such as a charge-coupled device (CCD) or a complementary metal-oxide semiconductor (CMOS), captures an environment outside the host vehicle 1 and includes at least luminance information. A luminance image (color image or monochrome image) can be generated. In addition, the imaging devices 110 are arranged in a substantially horizontal direction so that the optical axes of the two imaging devices 110 are substantially parallel on the traveling direction side of the host vehicle 1. The imaging device 110 continuously generates a luminance image obtained by imaging a three-dimensional object existing in the detection area in front of the host vehicle 1, for example, every 1/60 second frame (60 fps). Here, since the luminance images of different viewpoints are generated by the two imaging devices 110, the distance of the three-dimensional object can be grasped. Here, solid objects recognized by the imaging device 110 are only objects that exist independently such as vehicles (preceding vehicles, oncoming vehicles), pedestrians, street lights, traffic lights, roads (traveling paths), road signs, guardrails, and buildings. The thing which can be specified as a part of it is also included.

  The vehicle exterior environment recognition device 120 acquires a luminance image from each of the two imaging devices 110, and searches the other luminance image for a block corresponding to a block (an aggregate of a plurality of pixels) arbitrarily extracted from one luminance image. The so-called pattern matching is used to derive parallax (depth distance) and a screen position indicating the position of an arbitrary block in the screen, and a three-dimensional position of each block is derived. Then, the outside environment recognition device 120 identifies a three-dimensional object existing in the outside environment, for example, a preceding vehicle that travels in the same direction or an oncoming vehicle that travels oppositely. Further, when the three-dimensional object is specified in this way, the vehicle exterior environment recognition device 120 avoids a collision with the three-dimensional object (collision avoidance control), or keeps the distance between the vehicle and the preceding vehicle at a safe distance. Control (cruise control).

  In addition, the outside environment recognition device 120 receives a driver's request (intention) through the lighting switch 122 and performs light distribution control such as a headlamp through the lighting mechanism 124 according to the outside environment. As such light distribution control, for example, an HBA (High Beam Assist) that should not be irradiated with a high beam is turned off when a three-dimensional object such as a preceding vehicle or an oncoming vehicle is ahead, and turned on otherwise. If there is a solid object that should not be irradiated with a high beam when the area to be irradiated with the high beam is variable, other solid objects such as street lights, road signs, signboards, reflectors, etc. An ADB (Adaptive Driving Beam) that irradiates a high beam is an example of a region in which there is likely to exist. In order to realize such light distribution control, for example, as the illumination switch 122, a main switch that switches the position of the lamp to any one of the off, small lamp (position lamp), on (low beam), and auto lights. A dimmer switch is provided to switch the position between high beam impossible and high beam possible. In the case of HBA, the illumination mechanism 124 has a mechanism for switching between a low beam and a high beam, and in the case of ADB, the illumination mechanism 124 has a mechanism for changing the region of the high beam.

  Further, the vehicle control device 130 receives the driver's operation input through the steering wheel 132, the accelerator pedal 134, and the brake pedal 136, and transmits the operation input to the steering mechanism 142, the drive mechanism 144, and the brake mechanism 146 to control the host vehicle 1. . Further, the vehicle control device 130 controls the steering mechanism 142, the drive mechanism 144, and the braking mechanism 146 in accordance with instructions from the outside environment recognition device 120.

  As described above, in the outside environment recognition system 100, the preceding vehicle and the oncoming vehicle are quickly and accurately detected in order to cruise control the preceding vehicle and to avoid irradiating the preceding vehicle and the oncoming vehicle with a high beam. It is required to specify. In the outside environment recognition system 100, the information of the three-dimensional position and the color information are acquired through the luminance images obtained by the two imaging devices 110, so that the preceding vehicle and the oncoming vehicle can be quickly and accurately identified, and the headlamp is appropriately set. The purpose is to control light distribution.

  Hereinafter, the configuration of the outside environment recognition device 120 for realizing such an object will be described in detail. Here, the light distribution control of the headlamp, which is characteristic of the present embodiment, will be described in detail, and the description of the configuration unrelated to the characteristics of the present embodiment will be omitted.

(Vehicle environment recognition device 120)
FIG. 2 is a functional block diagram showing a schematic function of the outside environment recognition device 120. As shown in FIG. 2, the vehicle exterior environment recognition device 120 includes an I / F unit 150, a data holding unit 152, and a central control unit 154.

  The I / F unit 150 is an interface for performing bidirectional information exchange with the imaging device 110 and the vehicle control device 130. The data holding unit 152 includes a RAM, a flash memory, an HDD, and the like, and holds various pieces of information necessary for processing of each functional unit described below.

  The central control unit 154 is configured by a semiconductor integrated circuit including a central processing unit (CPU), a ROM storing a program, a RAM as a work area, and the like, and through the system bus 156, an I / F unit 150, a data holding unit 152 and the like are controlled. In the present embodiment, the central control unit 154 includes an image processing unit 160, a three-dimensional position deriving unit 162, a necessity determination unit 164, a detection range setting unit 166, a preceding vehicle extraction unit 168, a preceding vehicle recognition unit 170, It also functions as a vehicle extraction unit 172, a line-of-sight guidance light extraction unit 174, an oncoming vehicle recognition unit 176, a streetlight extraction unit 178, a streetlight recognition unit 180, a traveling scene determination unit 182 and a light distribution control unit 184. Hereinafter, the vehicle environment recognition processing for controlling the light distribution of the headlamp characteristic to the present embodiment will be described in detail based on the operation of each functional unit of the central control unit 154.

(External vehicle environment recognition processing)
FIG. 3 is a flowchart showing the flow of the external environment recognition process. In the outside environment recognition processing, the image processing unit 160 processes the image acquired from the imaging device 110 (S200), the three-dimensional position deriving unit 162 derives the three-dimensional position from the image (S202), and the necessity determining unit. 164 determines whether or not the high beam of the headlamp is unnecessary (S204). If it is determined that the high beam is not required (YES in S204), the vehicle exterior environment recognition process is terminated.

  If it is determined that the high beam is not necessary (NO in S204), the detection range setting unit 166 determines the detection ranges of the tail lamp, the headlamp, and the streetlight in the acquired image (S206), and extracts the preceding vehicle. The unit 168 extracts the tail lamp from the preceding vehicle detection range (S208), the preceding vehicle recognition unit 170 recognizes the preceding vehicle (S210), and the oncoming vehicle extraction unit 172 extracts the headlamp from the oncoming vehicle detection range. (S212) The oncoming vehicle recognition unit 176 recognizes the oncoming vehicle (S214), the streetlight extraction unit 178 extracts the streetlight from the streetlight detection range (S216), and the streetlight recognition unit 180 recognizes the streetlight (S218). ), The traveling scene determination unit 182 determines whether or not the traveling scene can be irradiated with a high beam from the position information of the streetlight (S220). , The light distribution control unit 184, a preceding vehicle, an oncoming vehicle, and performs the light distribution control of the headlamp on the basis of the traveling scene (S222), and finishes the environment outside the vehicle recognition processing. Hereinafter, each process is explained in full detail.

(Image processing S200)
The image processing unit 160 acquires a luminance image from each of the two imaging devices 110, and selects a block corresponding to a block arbitrarily extracted from one luminance image (for example, an array of horizontal 4 pixels × vertical 4 pixels) as the other luminance image. The parallax is derived using so-called pattern matching that is searched from the above. Here, “horizontal” indicates the horizontal direction of the screen, and “vertical” indicates the vertical direction of the screen.

  As this pattern matching, it is conceivable to compare the luminance (Y color difference signal) in units of blocks indicating an arbitrary image position between two luminance images. For example, SAD (Sum of Absolute Difference) that takes a luminance difference, SSD (Sum of Squared Intensity Difference) that uses the difference squared, or NCC that takes the similarity of a variance value obtained by subtracting an average value from the luminance of each pixel There are methods such as (Normalized Cross Correlation). The image processing unit 160 performs such a block-unit parallax derivation process for all blocks displayed in the detection area (for example, horizontal 600 pixels × vertical 180 pixels). Here, the block is assumed to be horizontal 4 pixels × vertical 4 pixels, but the number of pixels in the block can be arbitrarily set.

  However, the image processing unit 160 can derive the parallax for each block, which is a detection resolution unit, but cannot recognize what kind of three-dimensional object the block is. Therefore, the parallax is independently derived not in units of solid objects but in units of detection resolution (for example, blocks) in the detection region. Here, an image in which parallax information (corresponding to a depth distance z described later) derived in this way is associated with a luminance image is referred to as a distance image.

  FIG. 4 is an explanatory diagram for explaining the luminance image 212 and the distance image 214. For example, assume that a luminance image 212 as shown in FIG. 4A is generated for the image region 216 through the two imaging devices 110. However, here, for easy understanding, only one of the two luminance images 212 generated by each of the imaging devices 110 is schematically illustrated. In the present embodiment, the image processing unit 160 obtains the parallax for each block from such a luminance image 212 and forms a distance image 214 as shown in FIG. Each block in the distance image 214 is associated with the parallax of the block. Here, for convenience of description, blocks from which parallax is derived are represented by black dots.

(Three-dimensional position derivation process S202)
Subsequently, the three-dimensional position deriving unit 162 uses the so-called stereo method to calculate the parallax information for each block in the image area 216 based on the distance image 214 generated by the image processing unit 160, and the horizontal distance x, It is converted into a three-dimensional position in real space including depth y and depth distance (relative distance) z. Here, the stereo method is a method of deriving the depth distance z with respect to the imaging device 110 of the three-dimensional part from the parallax in the distance image 214 of the three-dimensional part (a pixel or a block composed of a plurality of pixels) by using a triangulation method. is there. At this time, the three-dimensional position deriving unit 162 is based on the depth distance z of the three-dimensional part and the detection distance on the distance image 214 between the point on the road surface at the same depth distance z as the three-dimensional part and the three-dimensional part. The height y of the three-dimensional part from the road surface is derived. Then, the derived three-dimensional position is associated with the distance image 214 again. Since various known techniques can be applied to the depth distance z deriving process and the three-dimensional position specifying process, description thereof is omitted here.

(Necessity determination processing S204)
Next, the necessity determination unit 164 determines whether the outside of the vehicle is bright (day or night), that is, whether the high beam of the headlamp is unnecessary. If the necessity determining unit 164 determines that the high beam is unnecessary (hereinafter, this state is simply referred to as a high beam unnecessary state), the subsequent headlamp light distribution control S206 to S222 is omitted.

  By the way, in the imaging device 110, an exposure adjustment unit (not shown) adjusts the exposure based on the environment outside the vehicle. Here, the exposure amount can be calculated based on the sensitivity (gain in this embodiment), the stop, and the exposure time. For example, the exposure adjustment unit uses the luminance distribution of a partial area (for example, a road surface area) of the generated image, and if the luminance of the area is high, the gain and the exposure time are reduced, and the luminance of the area is low. Increase the gain and exposure time. That is, the gain and the exposure time are adjusted so that the luminance is suitable for recognition of various three-dimensional objects.

  Therefore, the necessity determination unit 164 can grasp the external brightness by referring to the exposure amount adjusted by the exposure amount adjustment unit of the imaging device 110. For example, when the gain is small and the exposure time is short (when the exposure amount is small) compared to a predetermined threshold, it can be determined that the outside of the vehicle is bright and the high beam is unnecessary, and conversely, when the gain is large and the exposure time is long ( When the exposure amount is large, it can be determined that the outside of the vehicle is dark and the high beam can be used (hereinafter, this state is simply referred to as a high beam permission state).

(Detection range setting process S206)
When the necessity determination unit 164 determines that the high beam permission state is set (NO in S204), the detection range setting unit 166 detects the preceding vehicle (tail lamp), the oncoming vehicle (head lamp), and the streetlight in the acquired image. Determine the detection range. By limiting the detection range in the image in this way, it is possible to shorten the processing time and prevent erroneous detection in a region where the preceding vehicle and the oncoming vehicle do not originally exist.

  FIG. 5 is an explanatory diagram for explaining the detection range. The detection range setting unit 166 detects the preceding vehicle detection range 220a indicated by a broken-line rectangle at a predetermined position shown in FIG. 5 in the image area 216 for each of the preceding vehicle, the oncoming vehicle, and the streetlight to be detected. The oncoming vehicle detection range 220b indicated by the dashed-dotted rectangle and the streetlight detection range 220c indicated by the two-dot dashed rectangle are set. As can be understood with reference to FIG. 5, the preceding vehicle detection range 220a is included in the oncoming vehicle detection range 220b, and both and the streetlight detection range 220c are exclusive.

  The preceding vehicle detection range 220a, the oncoming vehicle detection range 220b, and the streetlight detection range 220c can be offset according to the environment outside the vehicle and the traveling path. For example, if the road is curved or has a slope, the preceding vehicle detection range 220a, the oncoming vehicle detection range 220b, and the streetlight detection range 220c are offset according to the degree. For example, if the travel path is a left curve, the detection range setting unit 166 offsets the preceding vehicle detection range 220a, the oncoming vehicle detection range 220b, and the streetlight detection range 220c to the left by an amount corresponding to the travel path. In this way, it is possible to set as a detection range a position where the possibility of the presence of a preceding vehicle, an oncoming vehicle, and a streetlight is highest.

(Preceding vehicle extraction process S208)
Subsequently, the preceding vehicle extraction unit 168 extracts a tail lamp from the preceding vehicle detection range 220a according to the luminance and color information and the three-dimensional position. However, the tail lamp of the preceding vehicle is different in light quantity from the head lamp and street lamp of the oncoming vehicle described later. Then, when the imaging device 110 captures images with an exposure time in which the tail lamp can be acquired, the brightness of the headlamps and street lamps is saturated, and conversely, if the imaging is performed with the exposure time in which the head lamps and street lights can be acquired, the tail lamp cannot be detected. End up. Therefore, the imaging apparatus 110 generates an image with at least two exposure times with different frames.

  FIG. 6 is an explanatory diagram for explaining luminance images 212 having different exposure times. For example, FIG. 6A shows a long exposure time, and FIG. 6B shows a short exposure time. Therefore, if the luminance image 212 in FIG. 6A is used, the luminance of the headlamp and the streetlight may be saturated, but the tail lamp can be appropriately extracted. If the luminance image 212 in FIG. 6B is used, the tail lamp is used. Although there is a possibility that it cannot be extracted, headlamps and street lamps can be appropriately extracted.

  In the preceding vehicle detection range 220a, the preceding vehicle extraction unit 168 has color information (RGB or YUV) within a predetermined color range indicating red, and a three-dimensional position within a predetermined distance range (for example, 1.5 pixels). ) Group the pixels in the parentheses. However, the preceding vehicle extraction unit 168 groups the pixels into a rectangular shape including a horizontal line and a vertical line that includes all the pixels that satisfy this condition. The grouped tail lamp candidates have basic feature amounts such as the vertical and horizontal coordinates of the group, the number of pixels in the group, the maximum luminance value in the group, the minimum luminance value, and the average depth distance (average parallax) of the group.

  Here, the preceding vehicle extraction unit 168 determines that the difference (size) between the vertical and horizontal coordinates of the group is a predetermined value (for example, two pixels) or less, and the difference (size) between the vertical and horizontal coordinates of the group is a predetermined value (determined by the depth distance). If any of the exclusion conditions such as the number of pixels in the group satisfying a predetermined value (for example, 2) or less is satisfied, it is excluded as a tail lamp candidate.

  By the way, a tail lamp and a head lamp (to be described below for the sake of brevity) will be extracted based on feature values (luminance and color information) in the image. If it fluctuates, the tail lamp extraction itself may become unstable. For example, for a preceding vehicle that actually exists, a situation is repeated in which the tail lamp is determined to be a tail lamp in an arbitrary frame but is not a tail lamp in another frame. When the tail lamp is extracted in an unstable manner as described above, hunting in which the light distribution control of the head lamp of the host vehicle 1 repeats a high beam and a low beam occurs.

  In addition, a three-dimensional object that does not emit light, such as a reflector located in the vicinity of the traveling path, has different image feature amounts depending on how the headlamp of the host vehicle 1 hits, and a tail lamp (preceding vehicle) or a headlamp (oncoming vehicle). ). For example, when the host vehicle 1 is set to a high beam, the reflection is misrecognized as a tail lamp (preceding vehicle), and the high beam is switched to a low beam so that the preceding vehicle is not irradiated with the high beam. However, when switching to the low beam, the reflection of the high beam disappears, and when it is not recognized as the tail lamp, hunting may occur such that the beam becomes a high beam again.

  For such hunting, countermeasure processing can also be executed in light distribution control after recognition of the preceding vehicle or oncoming vehicle, but if the recognition itself that is the basis of light distribution control is unstable, the light distribution control itself Must be complicated, resulting in loss of robustness.

  Therefore, in the present embodiment, at the time of extraction of the tail lamp or the like, processing in which hysteresis characteristics are incorporated into the recognition processing is performed. Specifically, the threshold value for comparing the feature amounts is varied depending on whether or not the target to be extracted is hit with a high beam.

  For example, in a region where a high beam is hit, a threshold value is set higher (stricter) than in a region where a high beam is not hit, thereby preventing erroneous detection of a tail lamp or the like. Alternatively, in a region where the high beam is not applied, the threshold is set lower (relaxed) than in a region where the high beam is applied, so that the tail lamp or the like can be easily extracted. In this way, tail lamps and the like can be appropriately extracted and hunting can be prevented.

  In this way, hunting of the high beam and the low beam can be prevented, and the possibility of hitting the high beam with respect to the preceding vehicle can be reduced.

(Preceding vehicle recognition process S210)
Subsequently, the preceding vehicle recognition unit 170 groups the tail lamps extracted by the preceding vehicle extraction unit 168 and recognizes the preceding vehicle in the preceding vehicle detection range 220a.

  Specifically, the distance on the image between the tail lamps (groups) is within a distance range (longer than the predetermined distance range in the preceding vehicle extraction process S208) included in the same vehicle, or a difference in average depth distance (average parallax) Are all within the distance range included in the same vehicle, or the ratio of the maximum luminance value is within a predetermined range, the tail lamps are grouped together as a preceding vehicle candidate.

  The preceding vehicle candidates grouped in this way inherit the basic feature values of the tail lamps before grouping. For example, the up / down / left / right coordinates of the preceding vehicle candidate inherit the up / down / left / right coordinates of the tail lamp corresponding to the outside of the preceding vehicle candidate, and the maximum brightness value and the minimum brightness value of the preceding vehicle candidate are the maximum brightness value and the minimum brightness value of the tail lamp. Both of these are inherited, and the average depth distance of the preceding vehicle candidate is inherited that the tail lamp has the shorter average depth distance (the one with the larger average parallax). Then, the number of tail lamps included in the preceding vehicle candidates is also counted.

  In addition, the preceding vehicle recognition unit 170 determines whether or not the presence of the preceding vehicle has been confirmed at an equivalent three-dimensional position in the past frame, and counts the number of existences that have been confirmed. The number of such times affects the reliability of the likelihood of the preceding vehicle. Then, the preceding vehicle recognition unit 170 determines whether the condition indicating that the preceding vehicle is reliable or the condition indicating that the preceding vehicle is not reliable is satisfied, and according to the result A preceding vehicle candidate is specified as a preceding vehicle, or excluded from a preceding vehicle candidate. Such a preceding vehicle recognition process S210 can employ various existing technologies such as Japanese Patent Application No. 2014-232431, and a detailed description thereof will be omitted here.

(Oncoming vehicle extraction process S212)
Subsequently, the oncoming vehicle extraction unit 172 extracts a headlamp from the oncoming vehicle detection range 220b according to the luminance and color information and the three-dimensional position. However, as described above, since the headlamp of the oncoming vehicle has a different light amount from the tail lamp of the preceding vehicle, the luminance image 212 having a short exposure time shown in FIG. 6B is used.

  Next, in the oncoming vehicle detection range 220b, the oncoming vehicle extraction unit 172 has a luminance that is equal to or higher than a predetermined luminance threshold (for example, 5 to 10 in 256 steps), and the three-dimensional position has a predetermined distance range (for example, 1.5). Pixels) are grouped together. However, the oncoming vehicle extraction unit 172 groups the pixels into a rectangular shape including a horizontal line and a vertical line that includes all the pixels that satisfy this condition. The grouped headlamp candidates have basic feature amounts such as the vertical and horizontal coordinates of the group, the number of pixels in the group, the maximum luminance value in the group, the minimum luminance value, the average depth distance (average parallax) of the group, and the YUV color average value.

  Here, the oncoming vehicle extraction unit 172 has a difference (size) between the vertical and horizontal coordinates of the group equal to or smaller than a predetermined value (for example, two pixels), and a difference (size) between the vertical and horizontal coordinates of the group is equal to or larger than a predetermined value (determined by the depth distance) If any of the exclusion conditions such that the number of pixels in the group satisfies a predetermined value (for example, 2) or less may be excluded as a headlamp candidate. Here, the predetermined value to be compared with the luminance is adjusted based on the predetermined value in the previous frame. Such an oncoming vehicle extraction process S212 can employ various existing technologies such as Japanese Patent Application No. 2014-232430, and a detailed description thereof will be omitted here.

  By the way, a high-luminance line-of-sight guide light (snow pole) may exist on the shoulder of the traveling path.

  FIG. 7 is an explanatory diagram for explaining an installation state of the line-of-sight guide lamp. The line-of-sight guide light is composed of, for example, a green LED, and is installed on the shoulder of the traveling path as shown by the broken line in FIG. 7 to clearly indicate the road edge or road alignment at the time of poor visibility, etc. It is intended to guide the eyes. If such a line-of-sight guide light is mistakenly extracted as a headlamp, and it is mistakenly recognized as an oncoming vehicle, control such as not irradiating the high-beam in the vicinity of the line-of-sight guide light that should originally be irradiated with a high beam works. End up.

  Therefore, the line-of-sight guide light extraction unit 174 determines whether or not the headlamp candidate is a line-of-sight guide light, and if it is determined that it is a line-of-sight guide light, it finally excludes it from the headlamp candidate. For example, the line-of-sight guide light extraction unit 174 recognizes whether or not the candidate is a line-of-sight guide lamp candidate (line-of-sight guide light candidate) based on the color characteristics (YUV color average value) of the block and the luminance distribution.

  FIG. 8 is an explanatory diagram for explaining the feature of the color. The line-of-sight guide lamp emits light by itself, and it is difficult to distinguish it from a headlamp only by luminance. However, the green LED of the line-of-sight guide lamp and the halogen lamp of the headlamp have the following color characteristics. For example, in the UV plan view of FIG. 8A, the green LED is represented by a region in which the blue difference U and the red difference V are both negative as surrounded by a broken line. The red color difference V also changes to a general line system. In contrast, the halogen lamp is represented by at least a region where the blue difference U is negative as surrounded by a solid line, and the red difference V hardly changes from 0 with respect to the change of the blue difference U.

  Further, in the YV plan view of FIG. 8B, the green LED is represented by at least a region where the red difference V is negative as surrounded by a broken line. Although the halogen lamp changes in the system, the red difference V does not change as much as the green LED with respect to the change in the luminance Y as surrounded by the solid line. However, in both the green LED and the halogen lamp, the blue difference U is not easily affected by the luminance Y. Therefore, in the YU plan view (not shown), the UV plan view and the YV plan view show the difference between the green LED and the halogen lamp. There is no difference.

  Then, the line-of-sight guide light extraction unit 174 provides an identification range representing the green LED region in FIGS. 8A and 8B, and if the color characteristics of the headlamp candidate are included in the identification range, It determines with green LED. Specifically, in the UV plan view of FIG. 8A, when the UV average value of the headlamp candidate is included in the identification range 230, it is determined that the headlamp candidate is a line-of-sight guide lamp candidate (green LED). Further, in the YV plan view of FIG. 8B, if the YV average value of the headlamp candidate is included in the identification range 232, it is determined that the headlamp candidate is a line-of-sight guide lamp candidate (green LED).

  However, if the depth distance of the oncoming vehicle is short, colors such as bonnets and bumpers are included in the relatively low-luminance area in the headlamp, and the red color of the entire headlamp candidate is actually a headlamp. The average value of the difference V becomes high and may be erroneously recognized as a line-of-sight guide light. Therefore, when the depth distance of the headlamp candidate is short, only the high luminance region is extracted to determine whether it is a line-of-sight guide light.

  FIG. 9 is an explanatory diagram for explaining the feature of the color. Here, when the headlamp candidate is present below a predetermined depth threshold (for example, 50 m), the luminance Y is within a predetermined high-luminance component range (for example, 256 levels of 50 or more and less than 100) among all the pixels of the headlamp candidate. ) Is extracted, and it is determined whether or not it is a line-of-sight guide lamp candidate based on the YV average value.

  Specifically, in the YV plan view of FIG. 9, if the average value of YV of the headlamp candidates is at the lower left of the identification line 234, that is, V <a × Y + b (a and b are constants indicating a primary line). For example, it is determined that the headlamp candidate is a line-of-sight guide lamp candidate (green LED). Here, the upper limit (for example, 100) is adopted for the high luminance component because if the luminance is too high, the red difference V may be saturated, and information on the red difference V cannot be effectively used.

  Subsequently, the line-of-sight guide light extraction unit 174 further identifies whether or not the headlamp candidate is determined to be a line-of-sight guide lamp candidate based on the color characteristics based on the luminance distribution. To do.

  FIG. 10 is an explanatory diagram for explaining a distant oncoming vehicle. When the oncoming vehicle is located far away, as shown in FIG. 10, the left and right headlamps provided on the oncoming vehicle and the reflective portions of the road surface may be combined and recognized as a light source in a single region. At this time, the bumper portion in the vicinity of the headlamp and the reflective portion of the road surface may have high luminance and a strong color component, and may show a color component as if it is a line-of-sight guide light. Therefore, the line-of-sight guide light extraction unit 174 determines whether the luminance distribution of the headlamp candidate indicates two light sources on the left and right or a single light source.

  FIG. 11 is an explanatory diagram illustrating the operation of the line-of-sight guide light extraction unit 174. Here, it is assumed that the luminance Y of each pixel of the headlamp candidate shown by a rectangle in FIG. 11A is distributed as shown in FIG. First, as shown in FIG. 11B, the line-of-sight guide light extraction unit 174 obtains the sum of luminance Y for each pixel column in the horizontal direction of the headlamp candidates. Therefore, the sum of the numbers corresponding to the pixels in the vertical direction is derived.

  Next, the line-of-sight guide light extraction unit 174 identifies the sum total that is the maximum value among the sum totals for the number of pixels in the vertical direction, and FIG. 11C illustrates the pixel row in the horizontal direction that maximizes the sum. Plot the coordinates and brightness. Then, the line-of-sight guide light extraction unit 174 smoothes the plotted points so as to eliminate noise and derives an approximate curve as shown in FIG. Subsequently, as shown in FIG. 11D, the line-of-sight guide light extraction unit 174 performs first-order differentiation on the approximate curve to obtain the number of zero-cross points from positive to negative, that is, the number of peaks of the approximate curve. Here, as shown in FIG. 11D, if there are two or more zero-cross points, it can be determined that the headlamp candidate is not a line-of-sight guide lamp candidate.

  Further, in FIG. 11 (f) obtained by linearly differentiating the approximate curve as shown in FIG. 11 (e) for other headlamp candidates, there is only one zero-cross point from positive to negative. It can be determined that it is a guide light candidate. At this time, the line-of-sight guide light extraction unit 174 counts the number of headlamp candidates determined to be line-of-sight guide light candidates for later processing.

  Subsequently, the line-of-sight guide light extraction unit 174 identifies the line-of-sight guide light candidates as line-of-sight guide lights based on the number and arrangement of the line-of-sight guide lamp candidates and excludes them from the head lamp candidates.

  In many cases, the line-of-sight guide lights are continuously arranged at equal heights at a plurality of equal intervals on the shoulder of the traveling path. Therefore, the line-of-sight guide light extraction unit 174 first determines that the line-of-sight guide lamp candidate is a line-of-sight guide lamp when the number of headlamp candidates determined to be line-of-sight guide lamp candidates is greater than or equal to a predetermined value (for example, 2). Identify. In this way, by setting the number limit, it is possible to avoid erroneously recognizing the headlamp as a line-of-sight guide lamp.

  Subsequently, the line-of-sight guide light extraction unit 174 re-determines headlamp candidates that have not become line-of-sight guide lamp candidates based on the specified arrangement of the line-of-sight guide lights.

  FIG. 12 is an explanatory diagram for explaining re-determination of the line-of-sight guide light. Specifically, on the plane of the horizontal distance x and the depth distance z shown in FIG. 12, the least square line 240 of the headlamp candidate identified as the line-of-sight guidance lamp, denoted by “x”, is derived. Then, the line-of-sight guidance light extraction unit 174 re-guides the line-of-sight guidance for the headlamp candidate indicated by “◯” in FIG. 12 whose distance from the least square line 240 is less than a predetermined distance threshold (for example, 2 pixels). Identify as a light. In this way, when the depth distance of the headlamp candidate is too close, saturation occurs, and color information cannot be acquired accurately, the light source that has not been properly determined as a candidate for the visual guidance light is appropriately identified as the visual guidance light. It becomes possible.

  Here, an example in which the headlamp candidate is re-determined on the plane of the horizontal distance x and the depth distance z has been described. However, the present invention is not limited to this, and the least-squares regarding the three-dimensional position including the height y is used. A straight line may be derived and re-determination may be performed. Further, the determination is not limited to the least square line, and may be a multiple-order approximate curve, or may be re-determined based on a plurality of least square lines.

  The line-of-sight guide light extraction unit 174 adds information indicating that it is a line-of-sight guide light to the headlamp candidate thus identified as the line-of-sight guide light, and excludes it from the headlamp candidate. For headlamp candidates that are not identified as line-of-sight guide lights, information indicating that they are headlamp candidates is maintained.

  With the above-described configuration, it is possible to distinguish between the headlamp and the line-of-sight guide lamp, and appropriately exclude the line-of-sight guide lamp from the headlamp candidates. However, with such determination alone, there may be a temporary loss of a gaze guidance light that cannot acquire a color (characteristic characteristic) at a distance or a gaze guidance light that cannot be seen due to the presence of a sharp curve or a preceding vehicle.

  For example, if the color characteristic of the line-of-sight guide light is too weak to reach the threshold value for determining the line-of-sight guide light, it should be recognized as a line-of-sight guide light, but it will be erroneously recognized as a headlamp. However, if the threshold is set to a low value, there is a high possibility that the headlamp will be erroneously recognized as a line-of-sight guide lamp. In addition, the gaze guidance light is originally intended to guide the driver's gaze in a dense fog or strong snowstorm, but under such circumstances, it is often difficult to extract the headlamp candidate itself, which may cause misrecognition. Increases nature.

  If the line-of-sight guide light cannot be appropriately extracted in this way, the number of headlamp candidates determined to be line-of-sight guide lights does not exceed the predetermined value, and the line-of-sight guide lamp candidate is identified as the line-of-sight guide light. It becomes impossible. If it does so, extraction of a gaze guidance lamp will become unstable and hunting of a high beam and a low beam may arise.

  Therefore, in the present embodiment, in the driving scene determination process S220, which is a process subsequent to the oncoming vehicle extraction process S212, it is determined whether or not the scene includes a line-of-sight guide light, and the scene includes a line-of-sight guide light. If there is, the existence area is specified.

  FIG. 13 is an explanatory diagram for explaining the existence region 248. When the previous frame is a scene where the line-of-sight guide light exists and the existence area 248 that is the area where the line-of-sight guide light exists is specified, the line of sight is extracted in the oncoming vehicle extraction process S212 in the next frame and thereafter. The line-of-sight guide light is appropriately extracted by using the result indicating that the guide light is present and the specified existence region 248. Therefore, here, it is determined whether or not it is determined that the scene includes the line-of-sight guide light in the frame before the previous time, and if it is determined that the scene includes the line-of-sight guide light, the existence area 248 is specified. Processing is performed on the assumption that

  As described above, in the oncoming vehicle extraction process S212, the oncoming vehicle extraction unit 172 groups pixels in which the luminance within the oncoming vehicle detection range 220b is equal to or higher than a predetermined luminance threshold (for example, 5 to 10 in 256 steps). The lamp candidate. Here, even if the brightness is not so high, it is extracted as a headlamp candidate.

  However, if it is known in advance that the line-of-sight guide light exists, and if the existence area 248 is grasped, the possibility that the line-of-sight guide light is located in the existence area 248 is high, and It is unlikely that the vehicle headlamp is located.

  The reason why the line-of-sight guide light is specified in the present embodiment is to exclude the line-of-sight guide light erroneously recognized as the head lamp candidate from the head lamp candidate. It is not necessary to extract it as a lamp candidate. Therefore, the oncoming vehicle extraction unit 172 sets the luminance threshold value in the specified existence area 248 higher than the area other than the existence area 248 if the scene has a line-of-sight guide light. Therefore, each pixel in the existence region 248 is not extracted as a headlamp candidate unless the luminance is somewhat higher than that of pixels in the region other than the existence region 248.

  With this configuration, even if the color characteristic of the low-intensity visual guide light in the existence area 248 does not reach the threshold value for determining the visual guide light, it is not extracted (misrecognized) as a headlamp candidate in the first place. The line-of-sight guide light can be appropriately excluded from the head lamp candidates.

  In addition, the line-of-sight guide lamp is originally intended to guide the driver's line of sight in a dense fog or strong snowstorm, but under such circumstances, the luminance is low and it is difficult to extract the headlamp candidate itself. Therefore, the oncoming vehicle extraction unit 172 may temporarily lower the brightness threshold in areas other than the specified existence area 248 if the scene includes a line-of-sight guide light. Therefore, even when the brightness of each pixel in the area other than the existence area 248 is reduced due to dense fog or strong snowstorm, it is possible to appropriately extract headlamp candidates.

  In addition, as described above, it is known in advance that the scene includes a line-of-sight guide light, and when the existence area 248 is recognized, the line-of-sight guide light is located in the existence area 248. The possibility is high. Therefore, the line-of-sight guide light extraction unit 174 extracts a line-of-sight guide light using the UV plan view of FIG. 8A or the YV plan view of FIG. To change.

  FIGS. 14A and 14B are explanatory diagrams for explaining the characteristics of colors. FIG. 14A shows a UV plan view, and FIG. 14B shows a YV plan view. Specifically, if it is a scene in which a line-of-sight guide light exists, the line-of-sight guide light extraction unit 174, as shown in FIGS. 14 (a) and 14 (b), for blocks in the specified existence area 248, The identification ranges 230 and 232 are expanded from the alternate long and short dash line to a solid line so that the area becomes, for example, twice or more.

  Accordingly, in the UV plan view of FIG. 14A, when the UV average value of the headlamp candidate in the existence region 248 is included in the expanded identification range 230, the headlamp candidate is a line-of-sight guide lamp candidate. judge. Further, in the YV plan view of FIG. 14B, if the YV average value of the headlamp candidates in the existence region 248 is included in the enlarged identification range 232, the headlamp candidate is a line-of-sight guide lamp candidate. Judge that there is.

  In this way, it becomes possible to appropriately recognize a light source that does not have strong color characteristics as a line-of-sight guide lamp in the existence region 248 as a line-of-sight guide lamp, and can appropriately exclude the line-of-sight guide lamp from headlamp candidates. .

(Oncoming vehicle recognition process S214)
Returning to the description of FIG. 3, the oncoming vehicle recognition unit 176 groups the headlamps extracted by the oncoming vehicle extraction unit 172, and recognizes the oncoming vehicle in the oncoming vehicle detection range 220b.

  Specifically, the distance on the image between the headlamps (groups) is in a distance range (longer than the predetermined distance range in the oncoming vehicle extraction process S212) included in the same vehicle, or the difference in average depth distance is the same. When all the conditions that the vehicle is within the distance range or the ratio of the maximum luminance value is within the predetermined range are satisfied, the headlamps are grouped to be the oncoming vehicle candidate.

  The oncoming vehicle candidates grouped in this way inherit the basic feature values of the headlamps before grouping. For example, the up / down / left / right coordinates of the oncoming vehicle candidate inherit the up / down / left / right coordinates of the headlamp corresponding to the outside of the oncoming vehicle candidate, and the maximum luminance value and the minimum luminance value of the oncoming vehicle candidate are the maximum luminance value and the minimum of the headlamp. All of the luminance values are inherited, and the average parallax (depth distance) of the oncoming vehicle candidate is inherited that has the shorter average depth distance of the headlamp (the one with the larger average parallax). The number of headlamps included in the oncoming vehicle candidate is also counted.

  Further, the oncoming vehicle recognition unit 176 determines whether or not the presence of the oncoming vehicle has been confirmed at an equivalent three-dimensional position in the past frame, and counts the number of times that the oncoming vehicle has been confirmed. The number of such presences affects the reliability of the oncoming vehicle quality. Then, the oncoming vehicle recognition unit 176 determines whether the condition indicating that the oncoming vehicle is reliable or the condition indicating that the oncoming vehicle is not reliable is satisfied, and according to the result The oncoming vehicle candidate is specified as an oncoming vehicle or excluded from the oncoming vehicle candidate. Such an oncoming vehicle recognition process S214 can employ various existing technologies such as Japanese Patent Application No. 2014-231301, and a detailed description thereof will be omitted here.

(Street lamp extraction process S216)
Subsequently, the streetlight extraction unit 178 extracts streetlights from the streetlight detection range 220c according to the luminance and color information and the three-dimensional position by the same processing as the oncoming vehicle extraction processing S212.

(Street lamp recognition process S218)
The streetlight recognition unit 180 recognizes the streetlight extracted by the streetlight extraction unit 178. Here, the streetlight is not a three-dimensional object that should not be irradiated with a high beam, but is used in the subsequent travel scene determination process S220.

(Running scene determination process S220)
The traveling scene determination unit 182 determines whether or not the traveling scene can be irradiated with a high beam. For example, when the vehicle speed is a predetermined value (for example, 20 km / h) or less, the traveling scene determination unit 182 determines that the scene does not require a high beam. Moreover, the traveling scene determination unit 182 determines that the high beam is unnecessary when the host vehicle 1 turns left or right. Furthermore, when the number of street lamps is greater than or equal to a predetermined number (for example, 3), the traveling scene determination unit 182 determines that the scene outside the vehicle is sufficiently bright and does not require a high beam. Such a traveling scene determination process S220 can employ various existing technologies such as Japanese Patent Application No. 2014-232408 and Japanese Patent Application No. 2014-232409, and therefore detailed description thereof will be omitted here.

  In addition, as described above, in the present embodiment, the traveling scene determination unit 182 determines whether or not it is a scene in which a line-of-sight guide light exists in addition to the above-described scene determination. This is because, in the oncoming vehicle extraction process S212 described above, the line-of-sight guide light is appropriately extracted based on the determination as to whether or not the scene includes the line-of-sight guide light. Here, the traveling scene determination unit 182 determines whether or not the plurality of line-of-sight guide lights extracted by the oncoming vehicle extraction unit 172 are distributed in a rectangular shape, and according to the continuity of the state distributed in the rectangular shape. It is determined whether the scene includes a line-of-sight guide light. And if it is a scene where a line-of-sight guide light exists, the existence area 248 is specified.

  FIG. 15 is an explanatory diagram for explaining the operation of the traveling scene determination unit 182. On the plane of the horizontal distance x and the depth distance z shown in FIG. 12, the least square line 250 of the line-of-sight guide light indicated by “x” extracted by the oncoming vehicle extraction unit 172 is derived. Then, the traveling scene determination unit 182 obtains the center of gravity 252 in the least square line 250 and derives a vertical line 254 that passes through the center of gravity 252 and is perpendicular to the least square line 250.

  Next, for each of the line-of-sight guide lights extracted by the oncoming vehicle extraction unit 172, the traveling scene determination unit 182 includes a distance d1 from the line-of-sight guide light to the vertical line 254, and a distance d2 from the line-of-sight guide light to the least square line. Is derived. Then, the traveling scene determination unit 182 determines whether or not the ratio (Σd1 / Σd2) of each of the distances d1 and d2 is equal to or greater than a predetermined ratio threshold (for example, 1.8). As a result, if it is equal to or greater than the ratio threshold value, it is determined that the plurality of line-of-sight guide lights are distributed in a rectangular shape, and if it is less than the ratio threshold value, it is determined that the plurality of line-of-sight guide lights are not distributed in a rectangular shape. .

  Next, the traveling scene determination unit 182 includes a line-of-sight guide lamp according to the continuity of the state distributed in a rectangular shape, for example, the number of times that the line-of-sight guide light is determined to be distributed in a rectangular shape. It is determined whether or not. Specifically, when it is determined that the line-of-sight guide light is distributed in a rectangular shape, the traveling scene determination unit 182 adds a predetermined value (for example, 1) to the integrated value, and the line-of-sight guide light is distributed in a rectangular shape. If it is determined that it is not, a predetermined value (for example, 1) is subtracted from the integrated value. Such an integrated value can take a value of 0-100. Therefore, when the integrated value reaches 100, no addition is performed, and when the integrated value becomes 0, no subtraction is performed.

  In addition, a hysteresis characteristic is provided in the threshold value for comparing the integrated values in order to avoid frequent reversal of the determination as to whether or not the scene has a line-of-sight guide light.

  FIG. 16 is an explanatory diagram showing an example of the determination process of the traveling scene determination unit 182. For example, the traveling scene determination unit 182 determines that the line of sight guide light exists when the integrated value is equal to or greater than a predetermined first integrated threshold (for example, 20), and the integrated value is smaller than the first integrated threshold. If it is less than 2 integration threshold (for example, 10), it is determined that the scene does not have a gaze guidance lamp.

  Thus, as shown in FIG. 16, when the integrated value becomes equal to or greater than the first integrated threshold and the traveling scene determination unit 182 determines that the scene is a scene in which a line-of-sight guide light is present, until the integrated value becomes less than the second integrated threshold. Is determined not to be a scene in which a line-of-sight guide light exists (determination is reversed). Further, if the integrated value is less than the second integrated threshold value and the traveling scene determination unit 182 determines that the scene is not a scene where the visual guide light exists, the visual guide light is not changed until the integrated value is equal to or greater than the first integrated threshold value. It is no longer determined that the scene exists.

  With this configuration, even if it is determined that the line-of-sight guide light does not exist temporarily, such as when the line-of-sight guide light overlaps with the preceding vehicle, even though the line-of-sight guide light exists, based on the past history, It is possible to determine that the scene stably includes the line-of-sight guide light.

  Thus, if it determines with it being a scene where a gaze guidance light exists, the driving | running | working scene determination part 182 will specify the presence area | region 248 where a gaze guidance light exists.

  FIG. 17 is an explanatory diagram showing an example of the determination process of the traveling scene determination unit 182. Specifically, the traveling scene determination unit 182 is indicated by a broken line separated from the least square line 250 illustrated in FIG. 15 by a predetermined distance for determining whether or not the plurality of line-of-sight guide lights are distributed in a rectangular shape. Parallel straight lines 256a and 256b are derived. A region sandwiched between the two parallel straight lines 256a and 256b is defined as a region 248 where the line-of-sight guide lamp exists. When coordinates are converted onto the luminance image 212, the result is as shown in FIG. As shown in FIG. 13, the presence area 248 is used in the oncoming vehicle extraction process S212 after the next frame.

  Note that the right end of the existence area 248 in the direction of the least square line 250 is set to a position extending to the right by a predetermined distance from the right end of the line-of-sight guide light extracted by the oncoming vehicle extraction unit 172 (taking a margin). The left end of 248 in the direction of the least square line 250 is set to a position extending to the left by a predetermined distance from the left end of the line-of-sight guide light extracted by the oncoming vehicle extraction unit 172 (with a margin).

  Here, it is desirable to irradiate a high beam in a scene where a line-of-sight guide lamp exists. Therefore, the traveling scene determination unit 182 determines that the scene where the line-of-sight guide light exists is a scene where the high beam can be used.

(Light distribution control process S222)
Finally, the light distribution control unit 184 performs light distribution control of the headlamps of the host vehicle 1 based on the preceding vehicle, the oncoming vehicle, and the traveling scene.

  FIG. 18 is an explanatory diagram showing the operation of the light distribution control unit 184. As shown in FIG. 18, when the traveling scene determination unit 182 determines that the scene does not require a high beam, it is an HBA or ADB, and a solid object that should not be irradiated (preceding vehicle, oncoming vehicle). Irrespective of the number, high beam irradiation is not performed. Also, if the traveling scene determination unit 182 determines that the scene permits the use of the high beam, in the case of the HBA, if there is one or more solid objects that should not be irradiated, the high beam should not be irradiated and should be irradiated. If there is no solid object, high beam irradiation is performed. In addition, in the case of ADB, if there is one or more solid objects that should not be irradiated, high beam irradiation should be performed on a part of the range while avoiding irradiation of the solid objects, and there should be no solid objects that should not be irradiated. For example, a high beam is irradiated over the entire range.

  FIG. 19 is an explanatory diagram for explaining ADB light distribution control. Here, if there are one or more solid objects that should not be irradiated in the ADB, as shown in FIG. 19A, the horizontal maximum width W of all the three-dimensional objects is calculated, and high beam irradiation is performed only outside the position. I do. However, if the ADB high beam can be divided and irradiated to the central portion, the maximum horizontal widths W1 and W2 of the preceding vehicle and the oncoming vehicle, which should not be irradiated, as shown in FIG. Calculate and irradiate the center and outside with a high beam.

  In the case of ADB, when the oncoming vehicle is very close to the host vehicle 1, it may be outside the oncoming vehicle detection range 220b. For this reason, when the oncoming vehicle approaches a certain depth distance (for example, 50 m), irradiation in a direction in which passing is expected is not performed for a certain period (for example, 1 sec).

  In such an ADB, the high beam irradiation range can be adjusted by switching the cut line angle. In addition, when the traveling scene determination unit 182 determines that the scene has a line-of-sight guide light, the cut line angle switching mode is changed.

  FIG. 20 is a top view for explaining the cut line angle. As shown in FIG. 20, when the cut line angle is determined, the irradiation range of the high beam is also determined. In addition, when recognizing a preceding vehicle or an oncoming vehicle, the light distribution control unit 184 changes the high beam irradiation range by switching the cut line angle. Here, in order to suppress a sudden change in the cut line angle, a low-pass filter (LPF) with a delay is provided for switching the cut line angle.

  By the way, as described above, it is desirable that the line-of-sight guide lamp originally emits a high beam. However, if the line-of-sight guide lamp is misrecognized as a headlamp, control is performed to avoid irradiation of the high beam to that region. . Therefore, if the recognition that the light source is a line-of-sight guide lamp becomes unstable even though the light source is a line-of-sight guide lamp, hunting between the high beam and the low beam may occur. Here, when it is known in advance that the line-of-sight guide light exists, it is highly likely that the line-of-sight guide lamp is located in the existence area 248, and even if a headlamp candidate is extracted, it is the line-of-sight guide light. There is a high possibility.

  Therefore, when the light distribution control unit 184 determines that the traveling scene determination unit 182 is a scene where the line-of-sight guide light exists, the time constant of the low-pass filter used for switching the cut line angle is further increased, and the line-of-sight guide light When the scene is determined not to exist, the cut line angle is delayed and switched. By doing so, even when the line-of-sight guide light is erroneously recognized as a headlamp candidate, it is possible to avoid a situation in which the cut line angle is changed in a complicated manner.

  The vehicle exterior environment recognition device 120 can appropriately execute the light distribution control of the headlamp.

  Also provided are a program for causing a computer to function as the vehicle environment recognition device 120, and a computer-readable storage medium such as a flexible disk, magneto-optical disk, ROM, CD, DVD, or BD on which the program is recorded. Here, the program refers to data processing means described in an arbitrary language or description method.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this embodiment. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Is done.

  For example, in the above-described embodiment, the light distribution control by HBA or ADB is described each time, but all the processes can be applied to any light distribution control.

  In the above-described embodiment, various threshold values are set, but such values can be changed as appropriate, and can be set based on experience values and experimental values.

  It should be noted that each step of the vehicle environment recognition processing in the present specification does not necessarily have to be processed in time series in the order described in the flowchart, and may include processing in parallel or by a subroutine.

  INDUSTRIAL APPLICABILITY The present invention can be used for an external environment recognition device that performs light distribution control of a headlamp according to the external environment.

120 Outside vehicle environment recognition device 124 Illumination mechanism 164 Necessity determining unit 166 Detection range setting unit 168 Leading vehicle extraction unit 170 Leading vehicle recognition unit 172 Oncoming vehicle extraction unit 174 Line-of-sight guidance light extraction unit 176 Oncoming vehicle recognition unit 178 Streetlight extraction unit 180 Streetlight Recognition unit 182 Travel scene determination unit 184 Light distribution control unit

Claims (2)

Computer
An oncoming vehicle extraction unit that extracts a headlamp candidate of the oncoming vehicle from the image, on the condition that at least the luminance of the pixel is equal to or higher than a predetermined luminance threshold;
A line-of-sight guide light extraction unit that extracts a line-of-sight guide light from the extracted head lamp candidates and excludes the line-of-sight guide light from the head lamp candidates;
An oncoming vehicle recognition unit for recognizing an oncoming vehicle based on the headlamp candidates;
Based on the extraction of the line-of-sight guide light, it is determined whether it is a scene where the line-of-sight guide light exists, and if it is determined that the line-of-sight guide light exists, a running scene determination that specifies an existing area where the line-of-sight guide light exists Part,
Function as
If the oncoming vehicle extraction unit determines that the scene has a line-of-sight guide light, the on-vehicle environment recognition device is characterized in that the brightness threshold value in the presence area is set higher than the area other than the presence area.   The on-vehicle environment recognition device according to claim 1, wherein the oncoming vehicle extraction unit lowers the luminance threshold in a region other than the presence region when it is determined that the scene includes a line-of-sight guide light.

Images