JP4151571B2 - In-vehicle obstacle detection device - Google Patents

Description

  The present invention relates to an on-vehicle obstacle detection device that detects an obstacle existing around a vehicle and notifies a driver of the obstacle.

  There is known an apparatus that detects an obstacle around the vehicle by sensing an image of the surroundings of the vehicle using a camera and detecting a detection target such as another vehicle from the image by pattern matching. For example, Patent Document 1).

JP 2000-148777 A

  In pattern matching, an image pattern that matches the feature to be detected is extracted from the image and recognized as an obstacle. For this reason, there is a problem that an erroneous detection occurs because an image in which a detection target is reflected on a mirror-like reflection object or a surrounding landscape that matches the characteristics of the detection target is erroneously extracted.

An on-vehicle obstacle detection device according to the present invention includes an imaging unit that acquires a captured image around the host vehicle, a detection unit that detects an obstacle around the host vehicle based on the captured image acquired by the imaging unit, When the obstacle is irradiated with light and a plurality of obstacles are detected by the detection means and the obstacles overlap each other on the captured image, the obstacle is farther from the own vehicle. A light projecting unit that changes an irradiation state of the light that is irradiated toward a position, and a plurality of light sources acquired by the imaging unit when the light irradiation unit changes the irradiation state of the light that irradiates the obstacle. And a determination unit that determines the reliability of the detection result of the obstacle by the detection unit based on the captured image.

  According to the present invention, an obstacle around the host vehicle is detected based on a captured image around the host vehicle acquired by the imaging unit, and a plurality of obstacles are detected, and the plurality of obstacles are mutually detected on the captured image. In the case of overlapping, the irradiation state of the light emitted from the light projecting means is changed toward the obstacle located farther from the own vehicle among the plurality of obstacles. Then, the reliability of the obstacle detection result is determined based on a plurality of captured images respectively acquired by the imaging unit when the light projecting unit changes the light irradiation state. Since it did in this way, it can be judged whether the detected obstacle is a true obstacle, and the obstacle which the detection target reflected in the mirror-like reflection object, and the obstacle It is possible to prevent false detection caused by surrounding scenery that matches the characteristics of the.

  FIG. 1 shows the configuration of an obstacle detection apparatus according to an embodiment of the present invention. This obstacle detection device is mounted on a vehicle, detects other vehicles in front of the vehicle as obstacles, and notifies the driver. The obstacle detection apparatus includes a camera 1, a projector 2, an image memory 3, a microcomputer 4, and a monitor 5.

  The camera 1 is installed at the front of the host vehicle so as to capture the front of the host vehicle, and outputs the captured image to the image memory 3 as an image signal. The amplification gain amount (hereinafter simply referred to as gain) of the output signal with respect to the input light amount in the camera 1 and the shutter speed and aperture of the camera 1 are controlled by the microcomputer 4 as described later. The camera 1 has light receiving sensitivity in a range including at least a wavelength band of light from the surroundings in a traveling environment of the vehicle and a wavelength band of light emitted from a projector 2 described later. As the camera 1, for example, a CCD camera or the like is used.

  The light projector 2 projects light forward of the vehicle under the control of the microcomputer 4. The range of light irradiated by the projector 2, the irradiation intensity, and the irradiation direction can be changed under the control of the microcomputer 4. For example, the projector 2 is configured so that light from a lamp light source is reflected to the front of the vehicle by a reflector. In this case, by adjusting the emission intensity of the lamp light source and the position of the reflector under the control of the microcomputer 4, the range of irradiation light, the irradiation intensity, and the irradiation direction can be changed. In addition, the wavelength band of the light irradiated by the projector 2 is set within the wavelength band in which the camera 1 has light receiving sensitivity.

  The image memory 3 is a frame memory for storing and holding the image signal input from the camera 1 as image data. The image data stored in the image memory 3 is output to the microcomputer 4 in units of image frames under the control of the microcomputer 4.

  The microcomputer 4 inputs image data from the image memory 3 and extracts obstacle (other vehicle) candidates from the image by pattern matching. As is well known, this pattern matching is an image processing method in which features of an object are set in advance according to brightness, shape, etc. in an image and a portion matching the features is extracted. Furthermore, when there are a plurality of extracted obstacle candidates, it is determined whether or not they overlap on the image. If they overlap, it is determined that they are farther from the vehicle in real space. Is a texture candidate. The texture means a picture drawn in a planar shape, and here represents that it is not a detection target of the present apparatus, that is, an actual vehicle. The distance in real space between the host vehicle and the extracted obstacle candidate can be determined by the position of the obstacle candidate on the image. It is approaching.

  Textures that cause erroneous detection occur when the vehicle is reflected in a mirror-like reflector or when the surrounding landscape matches the characteristics of the vehicle used for pattern matching. In order to prevent such erroneous detection due to texture, when a texture candidate is detected as described above, the projector 4 is controlled by the microcomputer 4 to project light toward the texture candidate, and the camera 1 is used. Take an image. Then, by comparing the captured images during light projection and non-light projection, it is determined whether the portion that is a texture candidate is actually a texture or another vehicle. As a result, if the texture is determined, it is excluded from the obstacle candidates. In this way, an actual obstacle is detected by excluding textures from the obstacle candidates extracted by pattern matching. The processing contents executed by the microcomputer 4 described above will be described in detail later.

  The monitor 5 displays the image of the obstacle detected by the microcomputer 4 as described above in the vehicle interior based on the image signal output from the microcomputer 4. The image display on the monitor 5 notifies the driver of the presence of an obstacle, that is, another vehicle in the vicinity. When an obstacle is displayed as an image on the monitor 5, the display form may be changed according to the distance from the host vehicle, or a different sound for each distance may be output simultaneously with the image display. By doing in this way, the approach degree of an obstacle and the own vehicle can be alert | reported to a driver more easily.

  FIG. 2 shows an example of attachment of the camera 1 and the projector 2 to the vehicle. FIG. 2 is a schematic diagram when the vehicle is viewed from the front. For example, the camera 1 is arranged on the rear surface of the rear-view mirror as shown in FIG. Other installation positions may be used as long as the front of the vehicle can be imaged. The projector 2 also serves as a headlamp of the vehicle and irradiates visible light forward of the vehicle. Thus, the projector 2 may also serve as a headlamp, or may be installed separately from the headlamp. In addition, since the shadow produced by the light projection from the projector 2 is imaged using the camera 1 as will be described later, the camera 1 and the projector 2 need to be installed at different positions.

  A state when a texture candidate is detected from obstacle candidates extracted by the pattern matching described above will be described with reference to FIGS. FIG. 3A is a diagram showing a state in which three other vehicles 200, 201, and 202 are traveling in front of the host vehicle 100. FIG. 2 is a diagram illustrating an example of an image captured from a host vehicle 100. FIG. When the other vehicles 200, 201, and 202 are extracted as obstacle candidates from the image of FIG. 3B by pattern matching, the other vehicle 200 and the other vehicle 201 that follows the vehicle overlap each other on the image. In this case, the distant other vehicle 200 is set as a texture candidate.

  FIG. 4A is a diagram illustrating a state in which the other vehicle 203 is traveling in front of the host vehicle 100 and the host vehicle 100 is reflected in the other vehicle 203. FIG. 4B is a diagram illustrating an image captured from the host vehicle 100 in FIG. From the image of FIG. 4B, a texture (represented as a texture 100 ′) resulting from the reflection of the other vehicle 203 and the host vehicle 100 is extracted as an obstacle candidate by pattern matching. Since the texture 100 ′ appears on the image so as to be far from the other vehicle 203, the texture 100 ′ is detected as a texture candidate.

  As described above, when a texture candidate is detected, this apparatus irradiates light toward the texture candidate using the projector 2, and determines whether or not the texture candidate is truly a texture. The method will be described below with reference to FIGS. FIG. 5 shows a case where, in the same manner as in FIG. 3, the distant other vehicle 200 among the other vehicles 200 and 201 in front of the host vehicle is detected as a texture candidate. In this case, as shown in FIG. 5A, when light is projected forward from the host vehicle 100, the light irradiated on the dotted line range is blocked by the other vehicle 201. Therefore, the shaded portion in the figure is behind the other vehicle 201, and light is not irradiated to that range. Thereby, the shadow of the other vehicle 201 appears on the other vehicle 200 which is a texture candidate.

  Here, as described above, since the camera 1 and the projector 2 are installed at different positions, the shadow of the other vehicle 201 that appears on the other vehicle 200 due to light irradiation from the projector 2 is imaged using the camera 1. can do. For example, as shown in FIG. 5A, light is projected from the right headlamp (light projector 2), and an image is captured using the camera 1 positioned at the left and right centers of the host vehicle 100. FIG. 5C shows an example of the projected image thus captured. FIG. 5B is an image example when light is not projected in the same vehicle position state. In this case, as shown in FIG. 5C, a shadow 300 of the other vehicle 201 appears on the other vehicle 200 in the captured image during the projection.

  FIG. 6 shows a case where the texture 100 ′ of the host vehicle reflected in the other vehicle 203 in front of the host vehicle is detected as a texture candidate, as in FIG. 4. At this time, when light is irradiated forward from the host vehicle 100 as shown in FIG. 6A, the amount of light irradiated to the dotted line range is blocked by the other vehicle 203, and the shaded range in the figure is the shadow of the other vehicle 203. Become. FIG. 6C shows an example of a projected image captured in this state. FIG. 6B is an example of an image when no light is projected in the same vehicle position. In this case, as shown in FIG. 6C, no shadow appears on the texture 100 ′ in the captured image being projected.

  As described above, when a texture candidate is an actual vehicle by irradiating light using the projector 2, a shadow appears in the texture candidate. On the other hand, when the texture is not a vehicle, no shadow appears in the texture candidate. The presence or absence of such a shadow can be confirmed by examining the luminance change of the texture candidate in the non-light-projected and captured images during the light projection. When a texture candidate has a portion with a small luminance change in the captured image, a shadow appears in that portion, and therefore it can be determined that the texture candidate is another vehicle. However, when the luminance changes uniformly, no shadow appears, so it can be determined that the texture candidate is a texture due to reflection and is not another vehicle.

  In the examples of FIGS. 5 and 6, light is projected from the right headlamp of the host vehicle, but either the right or left headlamp is set according to the positional relationship between the host vehicle and the texture candidate. You may make it select. At this time, as shown in FIGS. 5 and 6, when there is a texture candidate on the left side when viewed from the own vehicle, light is emitted from the right headlamp, and when located on the right side, the left headlamp is turned on. It is preferable to use it. By doing in this way, when it sees from the camera 1 with which the center of right and left was provided, the range of the shadow produced on a texture candidate by the other vehicle of this side can be enlarged.

  When an image for examining the brightness change of the texture candidate as described above is captured, it is preferable to change any of the gain, shutter speed, or aperture in the camera 1 as compared to normal imaging. The contents will be described with reference to FIG. FIG. 7 shows an example of the relationship between the brightness in real space and the luminance output of the camera 1. Assume that the camera 1 performs luminance output according to the brightness of the imaged object in accordance with one of the camera characteristics indicated by reference numerals 10 and 11.

  During normal imaging, that is, when imaging an image for extracting obstacle candidates by pattern matching, a wide range of brightness so that the brightness distribution of objects of various brightness in real space can be represented on the image The camera characteristics 10 that can cover the range are selected. On the other hand, when capturing an image for examining the luminance change of the texture candidate during light projection / non-light projection, the camera characteristic 11 is selected. The camera characteristic 11 is such that the luminance output largely changes around the brightness of the texture candidate indicated by reference numeral 12 when not projected. The brightness 12 at the time of non-projection can be known in advance from the luminance value of the image captured by the camera characteristic 10. As a result, a large amount of change in the brightness of the texture candidate on the image can be obtained by the change in brightness when light is emitted from the projector 2.

  For example, if the brightness of the texture candidate is changed by ΔB shown in the figure due to light projection from the projector 2, only a luminance change of ΔL1 occurs in the image when the camera characteristic is 10; A brightness change occurs. Thus, even if the brightness change is small, a large brightness change can be obtained in the image by changing the camera characteristics. As a result, the presence or absence of a shadow appearing in the texture candidate can be more accurately determined, and the detection accuracy of the other vehicle that is the detection target can be increased.

  The selection of the camera characteristics described above can be performed by changing any of the gain, shutter speed, or aperture of the camera 1. When the shutter speed or the aperture is changed, the amount of light input to the camera 1 can be adjusted by controlling the exposure amount during imaging. Further, when the gain is changed, the luminance output amount from the camera 1 with respect to the input light amount can be adjusted. Therefore, an arbitrary camera characteristic can be selected by controlling either one or both of the exposure amount and the gain. At this time, it is preferable that the camera characteristics 10 and 11 to be selected are not fixed and are changed according to the captured image.

  By performing the processing as described above, this apparatus detects other surrounding vehicles as obstacles based on the captured image, and determines whether the detected other vehicles are textures. Determine the reliability of the vehicle detection results. If it is not a texture, it is determined that the vehicle is reliable and other vehicles are displayed on the screen. If it is a texture, it is determined that the vehicle is not reliable and the other vehicles are not displayed on the screen. In this way, the detection result of the other vehicle is notified to the driver. A flowchart executed at this time is shown in FIG. The processing flow of FIG. 8 is executed by the microcomputer 4 at predetermined time intervals, for example, every imaging frame interval in the camera 1.

  In step S100, the image data recorded in the image memory 3 is read. In this image data, luminance information of an image captured by the camera 1 is represented. That is, an image captured by the camera 1 is read into the microcomputer 4 in step S100. In step S101, an image pattern for extracting an obstacle to be detected, that is, another vehicle, by pattern matching is read. This image pattern is data representing characteristics such as the brightness and shape of the vehicle, and is recorded in the microcomputer 4.

  In step S102, a portion matching the image pattern read in step S101 is extracted from the image read in step S100 using pattern matching. In step S103, the part extracted in step S102 is detected as an obstacle candidate. More specifically, in step S102, the degree of approximation with the image pattern for pattern matching is calculated for each region extracted from the read image, and in step S103, the calculated degree of approximation is a predetermined value. An area exceeding the threshold is detected as an obstacle candidate. Thus, obstacle candidates are detected by performing pattern matching in steps S102 and S103.

  In steps S104 and S105, in order to determine whether or not the obstacle candidates detected in step S103 overlap each other on the image, the position of each obstacle candidate on the image is obtained. For example, the coordinate position of the center point of the area occupied by each obstacle candidate in the image is obtained, and thereby the position of each obstacle candidate on the image is represented. In step S104, it is determined whether there is an obstacle candidate whose position on the image has not yet been obtained. If there is an obstacle candidate, the position on the image of any obstacle candidate is obtained in step S105. After execution of step S105, the process returns to step S104. If positions on the image have been obtained for all obstacle candidates, or if no obstacle candidates have been detected in step S103, a negative determination is made in step S104, and the process proceeds to step S106.

  In step S106, an overlap determination on the image of each obstacle candidate detected in step S103 is performed. At this time, based on the position on the image of each obstacle candidate obtained in step S105 and the size of the area occupied by each obstacle candidate in the image, whether or not the obstacle candidates overlap each other on the image. Determine whether. If it is determined that they overlap, a texture candidate that is farther away in the real space is used as a texture candidate. Note that the distance in real space from the host vehicle to each obstacle candidate can be determined by the position on the image as described above.

  In step S107, the obstacle candidate information detected as texture candidates in step S106 is listed and recorded as a texture candidate list. The information recorded here includes position information on each texture candidate image calculated in step S105. The texture candidate list is stored by the microcomputer 4, and the contents are updated each time new information is written.

  In step S108, the emission intensity map of the projector 2 stored in the microcomputer 4 is read. In this emission intensity map, the relationship between the distance from the host vehicle to the obstacle candidate and the emission intensity is recorded. In step S109, the intensity, range, and direction of light emitted from the projector 2 are set according to the position of the texture candidate in the real space. At this time, the distance and direction in the real space of each texture candidate are obtained from the position information on the image recorded in the texture candidate list in step S107. Then, the irradiation intensity is determined based on the calculated distance and the emission intensity map read in step S108, and the irradiation direction is determined based on the calculated direction. The irradiation range is determined so that the entire texture candidate can be irradiated based on the position and size of each texture candidate. When a plurality of texture candidates are detected, the irradiation range and irradiation direction from the projector 2 are determined so that all of them are irradiated with light. At this time, no light is projected from the projector 2 yet.

  In step S110, the brightness value of each texture candidate on the image is calculated. At this time, since the brightness in the image area of each texture candidate is usually not uniform, the average value or median value of the brightness in that area is used as the brightness value on the image of each texture candidate. In step S111, the gain and shutter speed of the camera 1 are controlled based on the luminance value on the texture candidate image calculated in step S110. At this time, as described with reference to FIG. 7, the brightness of each texture candidate is obtained from the calculated brightness value and the camera characteristics at that time, and the camera characteristics in which the brightness output changes greatly centering on the brightness are obtained. The gain and shutter speed of the camera 1 are controlled. Note that when a plurality of different brightness values are calculated for a plurality of texture candidates in step S110, the camera characteristics are in a range including all the brightness values indicating the respective brightness values.

  In step S <b> 112, an image at the time of non-light projection captured using the camera 1 is read from the image memory 3. In step S113, the projector 2 is turned on, and light is emitted from the projector 2 according to the conditions set in step S109. In this way, the light is emitted from the projector 2. In step S114, the captured image of the camera 1 when the light is projected from the projector 2 in step S113 is read from the image memory 3. In step S115, the projector 2 is turned off and the projection from the projector 2 is terminated.

  In step S116, the difference between the non-light-projected image read in step S112 and the light-projected image read in step S114 is calculated, and a difference image between them is obtained. In the difference image, the luminance change of each texture candidate before and after the light projection from the projector 2 is represented. In step S117, a region having a luminance value smaller than a predetermined value is extracted from the difference image obtained in step S116. This area is a part where the luminance is not changed by light projection, that is, a part where a shadow is generated.

  In step S118, it is determined whether or not each texture candidate is a texture by determining whether or not the area extracted in step S117, that is, a shadow portion, is included in the area on each texture candidate image. judge. If there is a texture candidate that does not include a shadow portion, it is determined that the texture candidate is a texture, and the process proceeds to step S119. In step S119, the obstacle determined in step S118 is excluded from the obstacle candidates detected in step S103. After all the textures are excluded in step S119, the process proceeds to step S120. If it is determined in step S118 that all texture candidates include a shadow portion and are not determined to be textures, the process proceeds from step S118 to step S120.

  In step S120, information on the obstacle candidates detected in step S103 is recorded as a detection list, except for those determined as textures in step S118 and excluded in step S119. The information recorded here is information for displaying the detection result image on the monitor 5. The detection list is stored by the microcomputer 4, and the contents are updated each time new information is written. In step S121, an image signal is output to the monitor 5 based on the obstacle candidate information in the detection list recorded in step S120. In this way, the detected obstacle, that is, the other vehicle is displayed on the monitor 5.

According to the embodiment described above, the following operational effects can be obtained.
(1) The other vehicle is detected as an obstacle candidate from the captured image in front of the host vehicle imaged using the camera 1, and light is emitted from the projector 2 to the texture candidate among the obstacle candidates. Then, based on the captured images acquired by the camera 1 during the non-projection from the projector 2 and during the projection, it is determined whether the texture candidate is a texture. It was decided to judge reliability. Since it did in this way, it can be judged whether the detected obstacle candidate is a real obstacle or a texture. As a result, it is possible to prevent erroneous detection caused by textures such as an image in which an obstacle to be detected is reflected on a mirror-like reflection object and surrounding scenery that matches the feature of the obstacle.

(2) The captured image when the projector 2 is not irradiating light and the captured image when the projector 2 is irradiating light are acquired by the camera 1, and a difference image between these captured images is obtained. To determine whether or not the texture candidate is a texture and determine the reliability of the detection result of the obstacle candidate. Specifically, based on the obtained difference image, it is determined whether or not a shadow portion appears in the captured image when the projector 2 is irradiating light. When a shadow portion appears, the texture candidate (obstacle candidate) is an obstacle, and the detection result is determined to be reliable. If it does not appear, the texture candidate (obstacle candidate) is a texture, and the detection result is determined to be unreliable. Since it did in this way, the reliability of the detection result of an obstacle candidate can be judged based on a captured image.

(3) When a plurality of obstacle candidates are detected and they overlap each other on the captured image, a texture candidate that is located farther from the vehicle in the real space is set as the texture candidate, Projection from the projector 2 was performed. Since it did in this way, the captured image for determining whether a texture candidate is a texture can be acquired using the camera 1. FIG.

(4) Since the intensity, range, and direction of light emitted by the projector 2 are set according to the position and size of the detected texture candidate (obstacle candidate), the projector 2 is directed toward the texture candidate. Can be used.

(5) Based on the captured image when the projector 2 is not irradiating light, the gain or exposure amount in the camera 1 when the projector 2 is irradiating light is controlled. Since it did in this way, even when there is little change of the brightness when light is irradiated from the light projector 2, a big luminance change can be obtained in a captured image. As a result, the presence or absence of a shadow appearing in the texture candidate can be more accurately determined, and the reliability determination accuracy of the obstacle candidate detection result can be increased.

(6) Projecting light by selectively using one of the light projectors 2 provided in pairs on the left and right of the host vehicle according to the positional relationship between the texture candidate (obstacle candidate) and the host vehicle. By doing so, it is possible to increase the range of shadows that appear on the texture candidates as viewed from the camera 1. Therefore, it is possible to easily determine the presence or absence of a shadow that appears in the texture candidate.

  In the above embodiment, an example has been described in which the front of the host vehicle is imaged using the camera 1 and another vehicle existing in front of the host vehicle is detected as an obstacle based on the captured image. However, the present invention is not limited to this content, and when other directions around the host vehicle, for example, the side or the rear, are imaged and an obstacle existing in that direction is detected based on the captured image. Is also applicable. Furthermore, the present invention can also be applied to the case where objects other than other vehicles, such as pedestrians, are detected as obstacles.

  In the above embodiment, either visible light or infrared light can be used as the light emitted by the projector 2. In the case of visible light, as described in the above embodiment, the projector 2 can be easily realized using the headlamp of the vehicle, and in the case of infrared light, The light can be projected without affecting the driving operation of other vehicles. Furthermore, the wavelength band of both visible light and infrared light may be included. When infrared light is used, the same effect can be obtained with either far infrared light or near infrared light.

  Furthermore, in the above embodiment, an example has been described in which the intensity, range, and direction of light emitted from the projector 2 are set according to the position and size of the texture candidate in the real space. However, as long as the texture candidate determination process as described above can be executed using the image captured by the camera 1 during the projection from the projector 2, these setting conditions for the light emitted from the projector 2 May be fixed. Alternatively, any one of these setting conditions may be changed and any of them may be fixed. In this case, the variable setting condition of the projector 2 can be changed according to one or both of the position and size of the texture candidate in the real space.

  Further, in the above embodiment, an example of determining the reliability of the detection result by acquiring a captured image when light is irradiated on an obstacle candidate detected during non-projection from the projector 2. Explained. However, the present invention can be similarly applied to a detection result during light projection by acquiring a captured image when the intensity of light to be irradiated is changed. In this case, the irradiation intensity of the changing light may be stronger or weaker than before the change. Thus, the present invention can be applied to, for example, a night vision apparatus that projects near-infrared light.

  In the above embodiment, the detection unit and the determination unit are realized by the microcomputer 4, but this is only an example, and each component is not limited to the above embodiment unless the characteristics of the present invention are impaired. .

It is a figure which shows the structure of the obstruction detection apparatus by one Embodiment of this invention. It is a figure which shows the example of attachment to the vehicle of a camera and the light projector. It is a figure for demonstrating a mode when a texture candidate is detected from an obstacle candidate, (a) is a mode that three other vehicles are drive | working ahead of the own vehicle, (b) is a figure of (a). Examples of images taken from the host vehicle are shown. It is a figure for demonstrating a mode when a texture candidate is detected from an obstacle candidate, (a) A mode that the own vehicle is reflected in the other vehicle which is drive | working ahead of the own vehicle, (b) is a figure. The example of an image imaged from the own vehicle at the time of (a) is each shown. It is a figure for demonstrating the method of irradiating light toward a texture candidate, and determining whether the texture candidate is a true texture, (a) is three other vehicles ahead of the own vehicle. (B) shows an example of a captured image when light is not projected in the state of (a), and (c) shows an example of a captured image when light is projected in the state of (a). ing. It is a figure for demonstrating the method of irradiating light toward a texture candidate, and determining whether the texture candidate is truly texture, (a) is the other vehicle which is drive | working ahead of the own vehicle (B) is an example of a captured image when light is not projected in the state of (a), and (c) is a captured image when light is projected in the state of (a). Each example is shown. It is a figure explaining the content which changes the gain and shutter speed of a camera compared with the time of normal imaging, when imaging the image for investigating the brightness change of a texture candidate. It is a flowchart performed in the obstacle detection apparatus by one Embodiment of this invention.

Explanation of symbols

1: Camera 2: Projector 3: Image memory 4: Microcomputer 5: Monitor

Claims (8)

Imaging means for acquiring a captured image around the host vehicle;
Detecting means for detecting obstacles around the host vehicle based on a captured image acquired by the imaging means;
When the obstacle is irradiated with light, a plurality of obstacles are detected by the detection means, and the obstacles overlap each other on the captured image, the host vehicle among the obstacles A light projecting means for changing the irradiation state of the light radiated toward what is located farther from
Reliability of the detection result of the obstacle by the detecting means based on a plurality of captured images respectively acquired by the imaging means when the light projecting means changes the irradiation state of the light irradiating the obstacle An on-vehicle obstacle detection device comprising: a determination means for determining The in-vehicle obstacle detection device according to claim 1,
The determination means includes a first captured image acquired when the light projecting means is irradiating light in the first irradiation state, and the light projecting means has at least light intensity as the first irradiation state. The vehicle-mounted obstacle characterized in that the reliability is determined based on a luminance change of the obstacle in a second captured image acquired when light is irradiated in a second irradiation state with different Detection device. The in-vehicle obstacle detection device according to claim 2,
The determination means determines whether or not a shadow portion appears in the luminance change of the obstacle. If it appears, the determination means determines that the detection result of the obstacle is reliable, and if it does not appear, the determination is reliable. A vehicle-mounted obstacle detection device characterized by determining that there is no sex. In the in-vehicle obstacle detection device according to claim 2 or 3,
An in-vehicle obstacle detection device that controls one or both of a gain and an exposure amount in the imaging unit when capturing the second captured image based on the first captured image. In the vehicle-mounted obstacle detection device according to any one of claims 1 to 4 ,
According to either or both of the position and size of the obstacle, at least one of the intensity, range or direction of light emitted when the light projecting means changes the light irradiation state is set. An in-vehicle obstacle detection device. In the vehicle obstacle detection device according to any one of claims 1 to 5 ,
The light projecting means is provided in pairs on the left and right of the host vehicle,
A vehicle-mounted obstacle detection device that selectively changes the light irradiation state by one of the pair of left and right light projecting means according to the positional relationship between the obstacle and the host vehicle. In the vehicle-mounted obstacle detection device according to any one of claims 1 to 6 ,
The vehicle-mounted obstacle detection device, wherein the light projecting means emits visible light. In the vehicle-mounted obstacle detection device according to any one of claims 1 to 6 ,
The vehicle-mounted obstacle detection device characterized in that the light projecting means emits infrared light.

Images