JP2010097410A - Vehicle detection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle detection device for detecting a preceding vehicle with high accuracy. <P>SOLUTION: The vehicle detection device 1 includes: a CCD camera 4 and a visible light image acquisition part 13a for acquiring a visible light image in front of a vehicle; a near-infrared light projector 5 for irradiating infrared light ahead of the vehicle; the CCD camera 4 and an near-infrared image acquisition part 15a for acquiring an infrared image in front of the vehicle irradiated with the infrared light by the near-infrared light projector 5; and a preceding vehicle detection part 16a for detecting the preceding vehicle based on the infrared image obtained by the CCD camera 4 and the near-infrared image acquisition part 15a, and the visible light image obtained by the CCD camera 4 and the visible light image acquisition part 13a. The vehicle detection device 1 compares the visible light image and the infrared image to distinguish between reflectors such as reflectors of the rear of the vehicle and light sources such as tail lamps, and recognizes shape thereof, and positional relation or the like to detect the preceding vehicle having the reflectors and the tail lamps with high accuracy. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a vehicle detection device that detects a preceding vehicle from image information obtained by imaging the front of the vehicle.

Conventionally, there is JP, 2006-244331, A as a technique of such a field. The night driving assistance device described in this publication takes an image in front of the host vehicle with a camera, detects a luminance region having a luminance equal to or higher than a predetermined value based on luminance information in the image, and luminance in the image The preceding vehicle is detected by detecting the tail lamp of the vehicle from the positional relationship of the areas.
JP 2006-244331 A

  However, the night driving assistance device described above has the following problems. In other words, it is difficult to clearly distinguish lighting such as buildings along the road from the tail lamp of the preceding vehicle only with the luminance information of the image in front of the host vehicle. There was a problem that the reliability of detection of the preceding vehicle was low.

  The present invention has been made to solve such a problem, and an object thereof is to provide a vehicle detection device capable of detecting a preceding vehicle with high accuracy.

  The present invention is a vehicle detection device that detects a preceding vehicle traveling in front of a vehicle, and includes visible light image acquisition means for acquiring a visible light image in front of the vehicle, and infrared projection that irradiates infrared light in front of the vehicle. Means, an infrared image acquisition means for acquiring an infrared image in front of the vehicle irradiated with infrared rays by the infrared projection means, an infrared image acquired by the infrared image acquisition means, and a visible light image acquired by the visible light image acquisition means And preceding vehicle detection means for detecting the preceding vehicle.

  In this vehicle detection device, a visible light image in front of the vehicle is acquired by the visible light image acquisition unit, and an infrared image in front of the vehicle irradiated with infrared rays from the infrared projection unit is acquired by the infrared image acquisition unit. In the visible light image, the light reflected from the headlights of the vehicle by the road reflector installed on the road and the reflector on the back of the preceding vehicle is reflected, but it is emitted from the light source provided in the building along the road and the tail lamp of the preceding vehicle. The dripping light appears more intense than the reflected light from reflectors. In the infrared image, the light reflected by the road reflector and the reflector on the back of the preceding vehicle is reflected strongly. On the other hand, the light source provided in the building along the road and the tail lamp of the preceding vehicle do not emit much infrared light. . Therefore, by comparing the visible light image and the infrared image, the reflector such as the reflector is distinguished from the light source such as the tail lamp, and the shape and the positional relationship are recognized, so that the preceding vehicle including the reflector and the tail lamp can be obtained. It can be detected with high accuracy.

  Further, the preceding vehicle detection means, when the difference in luminance value between the pixels in the infrared image acquired by the infrared image acquisition means and the visible light image acquired by the visible light image acquisition means is equal to or greater than the first threshold, Is preferably detected. The preceding vehicle detection means detects a reflector having a difference between the luminance of each pixel of the infrared image and the luminance of each pixel of the visible light image that is equal to or greater than a predetermined threshold, that is, a luminance that is high in the infrared image and low in the visible light image. Or a preceding vehicle having a reflector and a tail lamp based on a visible light image and an infrared image by detecting a region in the image in which a tail lamp having a low luminance in the infrared image and a high luminance in the visible light image is detected Is preferably realized.

  Further, the vehicle further includes an extraction unit that extracts a preceding vehicle candidate region based on at least one of the visible light image acquired by the visible light image acquisition unit and the infrared image acquired by the infrared image acquisition unit, and the preceding vehicle detection unit includes: It is preferable to detect the preceding vehicle based on the preceding vehicle candidate area extracted by the extracting means.

  In this vehicle detection device, the extraction means recognizes, for example, the luminance (brightness) of each pixel in the visible light image, and sets a pixel set as a tail lamp candidate from the shape and position of the pixel set having high luminance in the image. And the area around the tail lamp candidate is extracted as a preceding vehicle candidate area. Alternatively, the extracting means recognizes the brightness of each pixel in the infrared image, detects a reflector candidate on the back of the preceding vehicle from the shape and position of a pixel set having a high brightness in the image, and surrounds the reflector candidate Is extracted as a preceding vehicle candidate area. By previously extracting the preceding vehicle candidate area in this way, the preceding vehicle detection means detects in advance an object that is different from the shape and positional relationship of the tail lamp and reflector of the vehicle (that is, a light source and a reflector other than the vehicle) in the image. The preceding vehicle can be detected with higher accuracy.

  Moreover, it is preferable that an infrared light projection means switches from a light extinction state to a lighting state, when an extraction means extracts a preceding vehicle candidate area | region. In this case, until the preceding vehicle candidate area is extracted, the energy saving of the apparatus can be achieved by turning off the infrared light projecting unit.

  Moreover, it is preferable that an infrared light projection means switches from a lighting state to a light extinction state, when an extraction means extracts a preceding vehicle candidate area | region. In this case, it is possible to acquire the visible light image by turning off the infrared light projecting unit immediately after acquiring the infrared image by previously setting the infrared light projecting unit to the lighting state. As a result, there is no need to acquire an infrared image after waiting for the infrared irradiation amount to stabilize after the infrared light projecting means is turned on, and the difference in acquisition time between the infrared image and the visible light image can be reduced. Therefore, the speed of the preceding vehicle detection can be improved.

  In addition, the preceding vehicle detection unit is configured to detect the infrared ray when the luminance change amount per predetermined time in the infrared image acquired by the infrared image acquisition unit is equal to or less than the second threshold after the infrared projection unit is switched to the lighting state. It is preferable to detect the preceding vehicle based on the image. According to such a configuration, since the amount of infrared light irradiated by the infrared light projecting unit is stabilized, the luminance change amount per predetermined time in the infrared image settles below a predetermined threshold value, the infrared light projecting unit is turned on. It is possible to prevent an infrared image acquired in a state where the amount of infrared rays immediately after is unstable from being used for detection of a preceding vehicle, and to detect the preceding vehicle with higher accuracy.

  According to the present invention, the preceding vehicle can be detected with higher accuracy.

  Hereinafter, an embodiment of a vehicle detection device according to the present invention will be described with reference to the drawings.

  The vehicle detection device according to the present embodiment detects a preceding vehicle existing ahead of the host vehicle, and provides the detected preceding vehicle information to a driving support device such as an ACC [Adaptive Cruise Control] device. In the present embodiment, it is possible to detect a four-wheel preceding vehicle existing at a distance (for example, a distance 600 m to 700 m ahead of the host vehicle) that exceeds the detection range of a radar sensor or the like that detects an obstacle on the road at night. It is.

  With reference to FIGS. 1-8, the vehicle detection apparatus 1 which concerns on 1st Embodiment is demonstrated. FIG. 1 is a configuration diagram illustrating a vehicle detection device according to the first embodiment. FIG. 2 is a graph showing a wavelength region that can be imaged by the CCD camera shown in FIG. FIG. 3 is a diagram illustrating an example of a visible light image in front of the host vehicle. FIG. 4 is a diagram illustrating an example of a near-infrared image in front of the host vehicle. FIG. 5 is a diagram showing a difference in the preceding vehicle candidate region between the visible light image of FIG. 3 and the near-infrared image of FIG. FIG. 6 is a diagram illustrating another example of a visible light image in front of the host vehicle. FIG. 7 is a diagram illustrating another example of the near-infrared image in front of the host vehicle. FIG. 8 is a diagram showing a difference in the preceding vehicle candidate region between the visible light image of FIG. 6 and the near-infrared image of FIG. 3, 4, 6, and 7, the solid line indicates what is reflected in the image (for example, a taillight of a preceding vehicle that emits light or a reflector that reflects light). A broken line indicates what is present in the image (for example, a body of a preceding vehicle or a building along the road). 3 to 5 and FIGS. 6 to 8 differ only in the position and shape of the reflector of the preceding vehicle. The reflector 22 of the preceding vehicle 20 shown in FIGS. 3 to 5 has a circular shape and is provided below the tail lamp 21. The reflector 32 of the preceding vehicle 30 shown in FIGS. 6 to 8 has an annular shape and is provided to surround the tail lamp 31.

  The vehicle detection device 1 obtains a visible light image and a near infrared image in front of the host vehicle with a camera capable of imaging from a visible light wavelength region to a near infrared wavelength region, and between pixels in the visible light image and the near infrared image. The preceding vehicle 20 existing in front of the host vehicle is detected from the difference in brightness (brightness). For this purpose, the vehicle detection device 1 includes a CCD camera 4, a near-infrared projector 5, a driving support device 6, and a vehicle detection ECU [Electric Control Unit] 10. The vehicle detection ECU 10 includes a visible light image acquisition unit 13 a, a preceding unit. A vehicle candidate region extraction unit 14a, a near infrared image acquisition unit 15a, and a preceding vehicle detection unit 16a are configured.

  In the first embodiment, the near-infrared projector 5 corresponds to the infrared projector described in the claims, and the visible light image acquisition unit 13a includes the visible light described in the claims. The CCD camera 4 and the near-infrared image acquisition unit 15a correspond to the image acquisition unit, the near-infrared image acquisition unit described in the claims, and the preceding vehicle candidate area extraction unit 14a the extraction described in the claims. The preceding vehicle detection unit 16a corresponds to the preceding vehicle detection means.

  The CCD camera 4 is provided in front of the host vehicle (for example, on the windshield side of the rearview mirror) and captures an image in front of the host vehicle. The CCD camera 4 is provided with a light cut filter that limits the wavelength region of light, and picks up an image having a light wavelength of 600 nm to 1000 nm (from a visible red region to a near infrared region). The CCD camera 4 transmits the captured image ahead of the host vehicle to the vehicle detection ECU 10 as an image signal.

  The near-infrared projector 5 is provided on the front surface of the host vehicle (for example, in the headlight housing), and irradiates the front of the host vehicle with near-infrared rays (light having a wavelength region of 800 nm to 1000 nm). The near-infrared projector 5 irradiates near-infrared rays up to 600 m to 700 m ahead of the host vehicle. However, since the near-infrared projector 5 does not include visible light, the near-infrared projector 5 does not affect oncoming vehicles or the like. For the near infrared projector 5, for example, a halogen lamp is used. The near-infrared projector 5 is turned on by an ON signal transmitted from the vehicle detection ECU 10 and is turned off by an OFF signal transmitted from the vehicle detection ECU 10.

  The driving support device 6 is provided in the host vehicle and supports driving of the host vehicle by the driver. As such a driving support device 6, for example, an ACC [Adaptive Cruise Control] device that performs traveling control of the host vehicle following the recognized preceding vehicle, or highlighting the preceding vehicle on the display in the vehicle, is provided to the driver. There is a warning device that calls attention.

  The vehicle detection ECU 10 is an electronic control unit including a CPU [Central Processing Unit] that performs arithmetic processing, a ROM [Read Only Memory] and a RAM [Random Access Memory] that serve as a storage unit, an image processing chip, and the like. 4. It is electrically connected to the near-infrared projector 5 and the driving support device 6 to control the vehicle detection device 1 in an integrated manner. The vehicle detection ECU 10 loads an application program stored in the ROM into the RAM and executes it by the CPU, whereby the visible light image acquisition unit 13a, the preceding vehicle candidate region extraction unit 14a, the near infrared image acquisition unit 15a, the preceding vehicle detection The part 16a is configured.

  Based on the image signal transmitted from the CCD camera 4 with the near-infrared projector 5 turned off, the visible light image acquisition unit 13a is an image (hereinafter, referred to as a red region (600 nm to 800 nm) of visible light in front of the host vehicle. (Referred to as a visible light image A) (see FIGS. 3 and 6). In the visible light image A, light in a red region emitted from a light source such as the circular tail lamps 21 and 31 of the preceding vehicles 20 and 30 and the illumination 26 of a building (for example, a store or a house such as a convenience store) 25 along the road. It is clearly visible. The preceding vehicles 20, 30 and the building 25 shown in FIGS. 3 and 6 are all separated by an effective irradiation distance (for example, 100 m) of the headlight of the own vehicle, for example, the reflectors 22, 32 and the building 25 itself are not reflected in the visible light image A.

  Based on the visible light image A acquired by the visible light image acquisition unit 13a, the preceding vehicle candidate area extraction unit 14a extracts a preceding vehicle candidate area T, which is an area where it is estimated that a preceding vehicle exists in the image. Specifically, the preceding vehicle candidate area extraction unit 14a performs binarization processing for distinguishing each pixel in the visible light image A based on whether or not the luminance (brightness) of each pixel is brighter than a predetermined luminance threshold. (Processing in which a pixel whose luminance is equal to or higher than the luminance threshold is 1 and a pixel whose luminance is equal to or lower than the luminance threshold is 0) In the image, the tail lamps 21 and 31 of the preceding vehicles 20 and 30 appear as a pair of left and right light sources having the same area arranged in a substantially horizontal direction, so that a set of pixels having a luminance of 1 in the binarized image. A labeling process is performed on the area with respect to its shape, horizontal position and vertical position in the image, horizontal distance and area ratio between a plurality of pixel collection areas, and the like. Thereafter, two pixel sets estimated as the tail lamps 21 and 31 of the preceding vehicles 20 and 30 based on the result of the labeling process are detected as tail lamp candidates, and the pixel area of the tail lamp candidate and the rectangular area surrounding the periphery are preceded in the image. Extracted as a vehicle candidate area T. The size of the preceding vehicle candidate region T may be constant, but may be changed depending on the size of the area of the pixel set detected as a tail lamp candidate, or detected as a tail lamp candidate. The aspect may be changed in accordance with the horizontal distance between the two pixel sets (that is, the distance between the pair of left and right tail lamps 21 and 31).

  The near-infrared image acquisition unit 15a transmits an ON signal to the near-infrared projector 5 when the preceding vehicle candidate region extraction unit 14a extracts the preceding vehicle candidate region T. The near-infrared image acquisition unit 15a, based on the image signal transmitted from the CCD camera 4 in a state where the near-infrared projector 5 is turned on, an image from the red region of visible light to the near-infrared region in front of the host vehicle (hereinafter referred to as near-infrared region). (Referred to as image B). In this near-infrared image B, in addition to the tail lamps 21 and 31 of the preceding vehicles 20 and 30 and the illumination 26 of the building 25 along the road, the reflectors 22 and 30 provided below the tail lamps 21 and 31 in the preceding vehicles 20 and 30 are provided. The on-road reflector 28 provided on the guard rail 27 along the road 32 and the road reflects the near infrared rays and clearly appears (see FIGS. 4 and 7).

  Here, immediately after the near-infrared projector 5 is turned on, the amount of near-infrared rays to be irradiated is not stable, so that the reflected light from the reflectors 22 and 32 of the preceding vehicles 20 and 30 and the road reflector 28 of the guardrail 27 is also stable. Without being clearly reflected in the near-infrared image B. Therefore, when the near-infrared image acquisition unit 15a acquires the near-infrared image B, the preceding vehicle detection unit 16a has a unit time change amount of luminance of each pixel of the near-infrared image B (for example, a change amount of luminance per second). ) Is less than or equal to a predetermined luminance change threshold value (second threshold value). This luminance change threshold value is obtained by reducing the unit time change amount of the luminance of each pixel of the near infrared image B with time, that is, by reducing the unit time change amount of the near infrared reflected light irradiated by the near infrared projector 5. This value is used to determine whether or not the amount of near-infrared rays emitted by the projector 5 is stable, and the value is determined by experiments or the like.

  When the preceding vehicle detection unit 16a determines that the unit time change amount of the luminance of each pixel of the near-infrared image B is equal to or less than the luminance change threshold, the near-infrared light emitted by the near-infrared projector 5 is in a stable state. It is determined that the infrared image B has been acquired, and an OFF signal is transmitted to the near-infrared projector 5.

  When the preceding vehicle detection unit 16a transmits an OFF signal to the near-infrared projector 5, the luminance and near-infrared image in each pixel of the visible light image A within the preceding vehicle candidate region T extracted by the preceding vehicle candidate region extraction unit 14a. The difference with the brightness | luminance in each pixel of B is calculated, ie, the brightness | luminance of the reflector 22 which appeared in the image by irradiating a near infrared ray is recognized (refer FIG.5 and FIG.8). Specifically, as a difference in the preceding vehicle candidate area T between the visible light image A in FIG. 3 and the near infrared image B in FIG. 4 shown in FIG. 5, a circle provided below the tail lamp 21 in the preceding vehicle 20. A part of the reflected light of the shaped reflector 22 and the reflected light of the road reflector 28 is calculated. Further, as a difference in the preceding vehicle candidate area T between the visible light image A of FIG. 6 shown in FIG. 8 and the near-infrared image B of FIG. 7, an annular shape provided around the tail lamp 31 in the preceding vehicle 30. A part of the reflected light of the reflector 32 and the reflected light of the road reflector 28 is calculated.

  In the preceding vehicle candidate region T, the preceding vehicle detection unit 16a has a difference between the luminance in each pixel of the visible light image A and the luminance in each pixel of the near-infrared image B greater than a predetermined difference threshold (first threshold). Determine if there is something. This difference threshold value is used to remove the near-infrared reflected light other than the reflectors 22 and 32 (for example, the surface of the guard rail 27) as noise, and the value is determined by experiment.

  When the preceding vehicle detection unit 16a determines that there is a preceding vehicle candidate region T in which the difference between the luminance in each pixel of the visible light image A and the luminance in each pixel of the near-infrared image B is greater than a predetermined difference threshold, In the vehicle candidate region T, the preceding vehicle 20 is detected on the assumption that the preceding vehicle 20 including the tail lamps 21 and 31 and the reflectors 22 and 32 exists. The preceding vehicle detection unit 16a transmits the detected information of the preceding vehicles 20 and 30 to the driving support device 6 as preceding vehicle information.

  Next, with reference to FIG. 1 to FIG. 5 and FIG. 9, an operation flow in the vehicle detection device 1 will be described in detail. FIG. 9 is a flowchart showing a flow of operations of the vehicle detection device according to the first embodiment.

  In this vehicle detection device 1, the near-infrared projector 5 is turned off at the start of detection of the preceding vehicle. First, the CCD camera 4 captures an image in front of the host vehicle with the near-infrared projector 5 turned off, and transmits an image signal to the vehicle detection ECU 10. And the visible light image acquisition part 13a in vehicle detection ECU10 acquires the visible light image A ahead of the own vehicle based on the image signal transmitted from the CCD camera 4 in the state where the near-infrared projector 5 is turned off (S1). ).

  Based on the visible light image A acquired by the visible light image acquisition unit 13a, the preceding vehicle candidate area extraction unit 14a extracts a preceding vehicle candidate area T, which is an area where it is estimated that a preceding vehicle exists in the image (S2). ).

  In step S3, when the preceding vehicle candidate area T cannot be extracted, the preceding vehicle candidate area extraction unit 14a determines that there is no preceding vehicle and proceeds to step S11. On the other hand, when the preceding vehicle candidate area extraction unit 14a can extract the preceding vehicle candidate area T, the near infrared image acquisition unit 15a transmits an ON signal to the near infrared projector 5 (S4).

  The near-infrared projector 5 to which the ON signal has been transmitted is turned on and irradiates near-infrared rays in front of the host vehicle. The CCD camera 4 captures an image in front of the host vehicle with the near-infrared projector 5 turned on, and transmits an image signal to the vehicle detection ECU 10. The near infrared image acquisition unit 15a in the vehicle detection ECU 10 acquires a near infrared image B in front of the host vehicle based on the image signal transmitted from the CCD camera 4 (S5).

  Subsequently, the preceding vehicle detection unit 16a determines whether or not the unit time change amount of the luminance of each pixel of the near-infrared image B is equal to or less than a predetermined luminance change threshold value (S6). If the preceding vehicle detection unit 16a determines that the unit time change amount of luminance in each pixel of the near-infrared image B is not less than or equal to the luminance change threshold value, it determines that the unit time change amount of luminance is equal to or less than the luminance change threshold value. The above determination on the amount of change in luminance per unit time is repeated.

  When the preceding vehicle detection unit 16a determines that the unit time change amount of the luminance of each pixel of the near-infrared image B is equal to or less than the luminance change threshold value, the preceding vehicle detection unit 16a transmits an OFF signal to the near-infrared projector 5 (S7). The near-infrared projector 5 to which the OFF signal is transmitted is turned off. Thereafter, the preceding vehicle detection unit 16a calculates the difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B in the preceding vehicle candidate region T extracted by the preceding vehicle candidate region extraction unit 14a. Calculate (S8).

  In the preceding vehicle candidate area T, the preceding vehicle detection unit 16a determines whether or not the difference between the luminance in each pixel of the visible light image A and the luminance in each pixel of the near-infrared image B is larger than a predetermined difference threshold value. Is determined (S9). When the preceding vehicle detection unit 16a determines that there is no preceding vehicle candidate region T in which the difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B is greater than the difference threshold, there is a preceding vehicle. It judges that it does not carry out, and transfers to step S11.

  When the preceding vehicle detection unit 16a determines that there is a preceding vehicle candidate region T in which the difference between the luminance in each pixel of the visible light image A and the luminance in each pixel of the near-infrared image B is greater than the difference threshold, the preceding vehicle candidate In the region T, the preceding vehicle 20 is detected on the assumption that the preceding vehicle 20 including the tail lamp 21 and the reflector 22 exists. In step S <b> 11, the preceding vehicle detection unit 16 a transmits information regarding the detected preceding vehicle 20 or a determination that there is no preceding vehicle to the driving support device 6 as preceding vehicle information.

  According to the vehicle detection device 1 described above, the visible light image A in front of the own vehicle in a state where no near infrared light is irradiated from the near infrared light projector 5 and the front of the own vehicle in a state where the near infrared light is irradiated from the near infrared light projector 5. A near-infrared image B is acquired. In the visible light image A, there may be a case where the light reflected by the headlight of the own vehicle is reflected by the reflector 22 on the rear surface of the preceding vehicle 20 or the road reflector 28 provided on the guard rail 27 along the road. The light emitted from the illumination 26 of the building 25 and the tail lamp 21 of the preceding vehicle 20 appears stronger than the reflected light from the reflector 22 and the like. In the near-infrared image B, the light reflected by the reflector 22 and the road reflector 28 of the preceding vehicle 20 is reflected strongly, while the tail lamp 21 of the preceding vehicle 20 and the illumination 26 of the building 25 along the road are near-infrared. Therefore, there is not much difference in how the image is captured compared to the visible light image A. Therefore, the light source such as the tail lamp 21 and the reflector such as the reflector 22 are distinguished from each other by comparing the visible light image A and the near-infrared image B, and the reflector 22 and the tail lamp are recognized by recognizing the shape and positional relationship thereof. Therefore, it is possible to detect the preceding vehicle 20 provided with 21 with high accuracy. Since such a vehicle detection device 1 enables detection of a preceding vehicle based only on image information, the preceding vehicle existing at a distance away from the host vehicle beyond a detection range such as a radar sensor that detects an obstacle on the road. Can also be detected.

  Further, the preceding vehicle candidate area extraction unit 14a in the vehicle detection ECU 10 recognizes the luminance of each pixel of the visible light image A, and determines the tail lamp candidate of the preceding vehicle from the shape and position of a pixel set having high luminance in the image. And a peripheral region including the tail lamp candidate is extracted as a preceding vehicle candidate region T. By previously extracting the preceding vehicle candidate region T in this way, it is possible to exclude from the detection target in advance an object that greatly differs from the shape and positional relationship of the tail lamp of the vehicle in the image, and the preceding vehicle 20 is detected with higher accuracy. can do. Furthermore, since it is not necessary to calculate the difference in luminance of each pixel in a region other than the preceding vehicle candidate region T in the image, it is possible to reduce the processing load on the vehicle detection ECU 10.

  Further, in the preceding vehicle detection unit 16a in the vehicle detection ECU 10, the difference in luminance between pixels in the visible light image A acquired by the visible light image acquisition unit 13a and the near infrared image B acquired by the near infrared image acquisition unit 15a is predetermined. If it is equal to or greater than the difference threshold, the preceding vehicle is detected. Thereby, the preceding vehicle detection unit 16a is a reflector in which the difference between the luminance of the near-infrared image B and the luminance of the visible light image A is equal to or greater than the difference threshold, that is, the luminance is high in the near-infrared image B and low in the visible light image A. 22 can be detected. Furthermore, since the difference between the luminance of the near-infrared image B and the luminance of the visible light image A is equal to or less than the difference threshold, that is, the near-infrared reflected light from the surface of the guard rail 27 other than the reflector 22 can be removed as noise. Based on the light image A and the near-infrared image B, detection of the preceding vehicle 20 including the reflector 22 and the tail lamp 21 is suitably realized.

  Further, the near-infrared image acquisition unit 15a in the vehicle detection ECU 10 transmits an ON signal to the near-infrared projector 5 when the preceding vehicle candidate region extraction unit 14a extracts a preceding vehicle candidate region based on the visible light image A. The near-infrared projector 5 is switched from the unlit state to the lit state. Thus, the energy saving of the vehicle detection device 1 can be achieved by keeping the near-infrared projector 5 in the extinguished state until the preceding vehicle candidate region T is extracted.

  Further, in the preceding vehicle candidate area extraction unit 14a in the vehicle detection ECU 10, the unit time change in luminance in each pixel of the near-infrared image B acquired by the near-infrared image acquisition unit 15a after the near-infrared projector 5 is switched to the lighting state. When the amount is equal to or less than a predetermined luminance change threshold, the preceding vehicle is detected based on the near-infrared image B. According to such a configuration, the near-infrared image B acquired in an unstable state of the amount of near-infrared light immediately after the near-infrared projector 5 is turned on is prevented from being used for detection of the preceding vehicle. The vehicle 20 can be detected with higher accuracy.

  With reference to FIGS. 1-8, the vehicle detection apparatus 2 which concerns on 2nd Embodiment is demonstrated. In the vehicle detection apparatus 2, the same code | symbol is provided about the structure similar to the structure of the vehicle detection apparatus 1 which concerns on 1st Embodiment, and the description is abbreviate | omitted.

  The vehicle detection device 2 differs from the vehicle detection device 1 according to the first embodiment in the flow of operation and the configuration of the vehicle detection ECU 11, and acquires a near-infrared image before the visible light image. The preceding vehicle candidate area is extracted based on the infrared image. For this purpose, the vehicle detection device 2 includes a CCD camera 4, a near-infrared projector 5, a driving support device 6, and a vehicle detection ECU 11. The vehicle detection ECU 11 includes a visible light image acquisition unit 13b and a preceding vehicle candidate region extraction unit 14b. The near-infrared image acquisition part 15b and the preceding vehicle detection part 16b are comprised.

  In the second embodiment, the near-infrared projector 5 corresponds to the infrared projector described in the claims, and the CCD camera 4 and the visible light image acquisition unit 13b include the visible light described in the claims. The CCD camera 4 and the near-infrared image acquisition unit 15b correspond to the image acquisition means, the near-infrared image acquisition means described in the claims, and the preceding vehicle candidate area extraction unit 14b the extraction described in the claims. The preceding vehicle detection unit 16b corresponds to the preceding vehicle detection means.

  The visible light image acquisition unit 13b acquires the visible light image A based on the image signal transmitted from the CCD camera 4 with the near-infrared projector 5 turned off. In the visible light image A, light in a red region emitted from a light source such as the tail lamps 21 and 31 of the preceding vehicles 20 and 30 and the illumination 26 of the building 25 along the road is clearly shown (see FIGS. 3 and 6).

  The near-infrared image acquisition unit 15b transmits an ON signal to the near-infrared projector 5 at the start of detection of the preceding vehicle. When the near-infrared image acquisition unit 15 b transmits an ON signal to the near-infrared projector 5, the near-infrared image acquisition unit 15 b acquires the near-infrared image B based on the image signal transmitted from the CCD camera 4 with the near-infrared projector 5 turned on. In this near-infrared image B, in addition to the illumination 26 of the building 25 and the road reflector 28 of the guardrail 27, the tail lamps 21, 31 and the reflectors 22, 32 of the preceding vehicles 20, 30 are reflected (see FIGS. 4 and 7). .

  The near-infrared image acquisition unit 15b determines, for the acquired near-infrared image B, whether the unit time change amount of the luminance of each pixel of the near-infrared image B is equal to or less than a predetermined luminance change threshold (second threshold). To do. When the near-infrared image acquisition unit 15b determines that the unit-time change amount of the luminance of each pixel of the near-infrared image B is equal to or less than the luminance change threshold value, the near-infrared light amount emitted by the near-infrared projector 5 is stable. It is determined that the near-infrared image B has been acquired.

  If the preceding vehicle candidate region extraction unit 14b determines that the near infrared image acquisition unit 15b has acquired the stable near infrared image B, the preceding vehicle candidate region extraction unit 14b performs a preceding operation based on the near infrared image B acquired by the near infrared image acquisition unit 15b. A candidate vehicle region T is extracted. Specifically, the preceding vehicle candidate region extraction unit 14b performs binarization processing (luminance is equal to or greater than the luminance threshold value) for each pixel in the near-infrared image B depending on whether the luminance of each pixel is brighter than a predetermined luminance threshold value. 1) and a pixel whose luminance is equal to or lower than the luminance threshold value is 0). Since the tail lamps 21 and 31 and the reflectors 22 and 32 of the preceding vehicles 20 and 30 in the image are a pair of left and right reflectors having the same area, the luminance is increased in the image after the binarization process. A labeling process is performed on the collection area of one pixel with respect to its shape, horizontal position and vertical position in the image, horizontal distance and area ratio between a plurality of pixel collection areas, and the like. Thereafter, two pixel sets estimated as the tail lamps 21 and 31 of the preceding vehicles 20 and 30 based on the result of the labeling process are detected as tail lamp candidates, and two pixel sets estimated as the reflector 22 are detected as reflector candidates. Then, a tail lamp candidate pixel set or reflector candidate pixel set and a rectangular area surrounding the periphery are extracted as a preceding vehicle candidate area T in the image. The size of the preceding vehicle candidate region T may be constant, but may be changed depending on the size of the area of the pixel set detected as a tail lamp candidate or a reflector candidate, for example. The preceding vehicle candidate area extraction unit 14b transmits an OFF signal to the near-infrared projector 5 when the preceding vehicle candidate area T is extracted.

  The preceding vehicle detection unit 16b calculates a difference between the luminance at each pixel of the near-infrared image B and the luminance at each pixel of the visible light image A in the preceding vehicle candidate region T extracted by the preceding vehicle candidate region extraction unit 14b. That is, the brightness of the reflectors 22 and 32 that are no longer visible in the image is recognized by turning off the near-infrared projector 5 (see FIGS. 5 and 8).

  In the preceding vehicle candidate region T, the preceding vehicle detection unit 16b has a difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B larger than a predetermined difference threshold (first threshold). Determine if there is something.

  When the preceding vehicle detection unit 16b determines that there is a preceding vehicle candidate region T in which the difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B is greater than the difference threshold, the preceding vehicle candidate In the region T, the preceding vehicles 20 and 30 are detected on the assumption that the preceding vehicles 20 and 30 having the tail lamps 21 and 31 and the reflectors 22 and 32 exist. The preceding vehicle detection unit 16b transmits information on the preceding vehicle to the driving support device 6 as preceding vehicle information.

  Next, with reference to FIG. 1 to FIG. 5 and FIG. 10, the operation flow in the vehicle detection device 2 will be described in detail. FIG. 10 is a flowchart showing an operation flow in the vehicle detection device according to the second embodiment.

  The near-infrared image acquisition unit 15b in the vehicle detection ECU 11 transmits an ON signal to the near-infrared projector 5 at the start of detection of the preceding vehicle (S21). The near-infrared projector 5 to which the ON signal is transmitted enters a lighting state and irradiates near-infrared rays in front of the host vehicle. The CCD camera 4 images the front of the host vehicle with the near-infrared projector 5 turned on. The near infrared image acquisition unit 15b acquires a near infrared image B in front of the host vehicle based on the image signal transmitted from the CCD camera 4 (S22).

  Next, the near-infrared image acquisition part 15b determines whether the unit time variation | change_quantity of the brightness | luminance of each pixel of the near-infrared image B is below a brightness | luminance change threshold value (S23). If the near-infrared image acquisition unit 15b determines that the unit time change amount of the luminance of each pixel of the near-infrared image B is not less than the luminance change threshold value, the near-infrared image acquisition unit 15b determines that the unit time change amount of the luminance is less than the luminance change threshold value. The determination regarding the unit time variation of the luminance is repeated until the above.

  When the near-infrared image acquisition unit 15b determines that the unit-time change amount of the luminance of each pixel of the near-infrared image B is equal to or less than the luminance change threshold value, the near-infrared light amount emitted by the near-infrared projector 5 is stable. It is determined that the near-infrared image B has been acquired. If the preceding vehicle candidate region extraction unit 14b determines that the near infrared image acquisition unit 15b has acquired the stable near infrared image B, the preceding vehicle candidate region extraction unit 14b performs a preceding operation based on the near infrared image B acquired by the near infrared image acquisition unit 15b. A candidate vehicle region T is extracted (S24).

  In step S25, when the preceding vehicle candidate area extraction unit 14b cannot extract the preceding vehicle candidate area T, the preceding vehicle candidate area extracting unit 14b determines that there is no preceding vehicle and proceeds to step S31. If the preceding vehicle candidate area extraction unit 14b can extract the preceding vehicle candidate area T, it transmits an OFF signal to the near-infrared projector 5 (S26). The near-infrared projector 5 to which the OFF signal is transmitted is turned off. The CCD camera 4 images the front of the host vehicle with the near-infrared projector 5 turned off. The visible light image acquisition unit 13b acquires the visible light image A in front of the host vehicle based on the image signal transmitted from the CCD camera 4 with the near-infrared projector 5 turned off (S27).

  Subsequently, the preceding vehicle detection unit 16b determines the difference between the luminance at each pixel of the near-infrared image B and the luminance at each pixel of the visible light image A in the preceding vehicle candidate region T extracted by the preceding vehicle candidate region extraction unit 14b. Is calculated (S28).

  Thereafter, in the preceding vehicle candidate area T, the preceding vehicle detection unit 16b has a difference between the luminance at each pixel of the near-infrared image B and the luminance at each pixel of the visible light image A greater than a predetermined difference threshold value. It is determined whether or not (S29). When the preceding vehicle detection unit 16b determines that there is no preceding vehicle candidate region T in which the difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B is greater than the difference threshold, there is a preceding vehicle. It judges that it does not carry out, and transfers to step S31.

  When the preceding vehicle detection unit 16b determines that there is a preceding vehicle candidate region T in which the difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B is greater than the difference threshold, the preceding vehicle candidate The preceding vehicle 20 is detected as the preceding vehicle 20 including the tail lamp 21 and the reflector 22 exists in the region T (S30). In step S <b> 31, the preceding vehicle detection unit 16 b transmits information regarding the detected preceding vehicle 20 or a determination that there is no preceding vehicle to the driving support device 6 as preceding vehicle information.

  The vehicle detection device 2 according to the second embodiment described above has the same effects as the vehicle detection device 1 according to the first embodiment, but has the following effects. That is, according to the vehicle detection device 2, a near-infrared image is acquired first, and after the preceding vehicle candidate area T is extracted from the near-infrared image, the near-infrared projector 5 is turned off and a visible light image is acquired. Since it is a structure, it is not necessary to wait for the near-infrared light quantity which the near-infrared light projector 5 irradiates after the preceding vehicle candidate area | region T is extracted to be stabilized. As a result, it is possible to shorten the time from when the preceding vehicle candidate area T is extracted until the preceding vehicle is detected, and the speed of detecting the preceding vehicle can be improved.

  A vehicle detection device 3 according to a third embodiment will be described with reference to FIGS. 3, 4, 6, 7, and 11 to 14. In the vehicle detection apparatus 3, the same code | symbol is provided about the structure similar to the structure of the vehicle detection apparatus 1 which concerns on 1st Embodiment, and the description is abbreviate | omitted. FIG. 11 is a configuration diagram illustrating a vehicle detection device according to the third embodiment. 12A is a graph showing the wavelength region that can be imaged by the visible light camera shown in FIG. 11, and FIG. 12B is a graph showing the wavelength region that can be imaged by the near-infrared camera shown in FIG. is there. FIG. 13 is a diagram illustrating a difference in the preceding vehicle candidate region between the visible light image of FIG. 3 and the near-infrared image of FIG. 4 according to the third embodiment. FIG. 14 is a diagram illustrating a difference in the preceding vehicle candidate region between the visible light image of FIG. 6 and the near-infrared image of FIG. 7 according to the third embodiment.

  Compared to the vehicle detection device 1 according to the first embodiment, the vehicle detection device 3 includes a visible light camera 7 and a near-infrared camera 8 instead of the CCD camera 4, and a vehicle detection ECU 12. The function of the preceding vehicle detection unit 17 is different. In the vehicle detection device 3, the visible light camera 7 captures a visible light image in front of the host vehicle, and the near-infrared camera 8 captures a near-infrared image in front of the host vehicle. A preceding vehicle is detected from the infrared image. For this purpose, the vehicle detection device 3 includes a near-infrared projector 5, a driving support device 6, a visible light camera 7, a near-infrared camera 8, and a vehicle detection ECU 12. The vehicle detection ECU 12 includes a visible light image acquisition unit 13a, a preceding unit. A vehicle candidate area extraction unit 14a, a near infrared image acquisition unit 15a, and a preceding vehicle detection unit 17 are configured.

  In the third embodiment, the near-infrared projector 5 corresponds to the infrared projector described in the claims, and the visible light camera 7 and the visible light image acquisition unit 13a are visible in the claims. The near-infrared camera 8 and the near-infrared image acquisition unit 15a correspond to the optical image acquisition unit, the near-infrared image acquisition unit described in the claims, and the preceding vehicle candidate area extraction unit 14a described in the claims. The preceding vehicle detection unit 16a corresponds to the preceding vehicle detection means.

  Hereinafter, the visible light camera 7, the near infrared camera 8, and the preceding vehicle detection unit 17 in the vehicle detection ECU 12 different from the first embodiment will be described in detail.

  The visible light camera 7 is a CCD camera that is provided in front of the host vehicle and captures an image in front of the host vehicle. The visible light camera 7 captures an image having a light wavelength of 400 nm to 1000 nm (visible light wavelength region) (see FIG. 12A). The visible light camera 7 transmits the captured image ahead of the host vehicle to the vehicle detection ECU 12 as a visible light image signal. The visible light image acquisition unit 13 a in the vehicle detection ECU 12 acquires the visible light image A from the visible light image signal transmitted by the visible light camera 7. In the visible light image A acquired from the visible light camera 7, light emitted from a light source such as the tail lamps 21 and 31 of the preceding vehicles 20 and 30 and the illumination 26 of the building 25 along the road is clearly shown (FIGS. 3 and 3). 6).

  The near-infrared camera 8 is a CCD camera that is provided in front of the host vehicle and captures an image in front of the host vehicle. The near-infrared camera 8 is provided with a light cut filter that limits the wavelength region of light, and captures an image with a light wavelength of 800 nm to 1000 nm (near-infrared wavelength region) (see FIG. 12B). The near-infrared camera 8 transmits, to the vehicle detection ECU 12, a front image of the host vehicle imaged with the near-infrared projector 5 turned on as a near-infrared image signal. The near infrared image acquisition unit 15a in the vehicle detection ECU 12 acquires the near infrared image B from the near infrared image signal transmitted by the near infrared camera 8. In the near-infrared image B acquired from the near-infrared camera 8, light sources such as the tail lamps 21 and 31 of the preceding vehicles 20 and 30 and the illumination 26 of the building 25 along the road hardly emit light in the near-infrared wavelength region. Therefore, while it appears weak, the on-road reflector 28 provided on the reflectors 22 and 32 of the preceding vehicles 20 and 30 and the guard rail 27 along the road reflects near infrared rays and clearly appears (see FIGS. 4 and 7). In the near-infrared image B, since visible light is not captured, the brightness of the tail lamps 21 and 31 and the illumination 26 that hardly emit near-infrared light are lower than those in the first and second embodiments.

  In the preceding vehicle candidate region T extracted in the same manner as in the first embodiment or the second embodiment, the preceding vehicle detection unit 17 determines the luminance in each pixel of the visible light image A and each of the near infrared images B. The difference with the luminance at the pixel is calculated. Here, the light emitted from the tail lamps 21 and 31 of the preceding vehicles 20 and 30 and the illumination 26 of the building 25 appears strong in the visible light image A, but weakly appears in the near-infrared image B because it does not contain much near-infrared light. Further, the reflectors 22 and 32 of the preceding vehicles 20 and 30 and the road reflector 28 of the guard rail 27 are hardly reflected in the visible light image A because they are more than the effective irradiation distance of the headlight of the own vehicle, but the near infrared image. In B, the light reflected by the near-infrared projector 5 is reflected strongly.

  Therefore, the preceding vehicle detection unit 17 calculates the difference between the luminance of each pixel of the visible light image A and the luminance of each pixel of the near-infrared image B in the preceding vehicle candidate area T, thereby calculating the preceding vehicles 20 and 30. The tail lamps 21 and 31 and the reflectors 22 and 32 are recognized (see FIGS. 13 and 14). Specifically, as the difference in the preceding vehicle candidate area T between the visible light image A of FIG. 3 and the near infrared image B of FIG. 4 shown in FIG. A part of the reflected light of the circular reflector 22 and the reflected light of the road reflector 28 is calculated, and the visible light image A of FIG. 6 and the near-infrared image B of FIG. 7 shown in FIG. As the difference in the vehicle candidate region T, the light of the tail lamp 31 of the preceding vehicle 30, the reflected light of the annular reflector 32 provided around the tail lamp 31, and a part of the reflected light of the road reflector 28 are calculated. The

  Thereafter, the preceding vehicle detection unit 17 determines that there is a difference between the luminance of each pixel of the visible light image A and the luminance of each pixel of the near-infrared image B in the preceding vehicle candidate region T being larger than a predetermined difference threshold. In this case, the difference between the luminance of each pixel of the visible light image A and the luminance of each pixel of the near-infrared image B is recognized, and the tail lamps 21 and 31 and the reflectors 22 and 32 are distinguished. Specifically, when the luminance of each pixel in the near-infrared image B is subtracted from the luminance of each pixel in the visible light image A, a pixel set having a positive difference value (that is, the visible light image A has a high luminance and a near infrared ray). A pixel set having a low luminance in the image B) is determined as a light source such as the tail lamps 21 and 31, and a pixel set having a negative difference value (that is, a pixel set having a low luminance in the visible light image A and a high luminance in the near-infrared image B). Is a reflector such as the reflectors 22 and 32. In this manner, the preceding vehicle detection unit 17 detects the preceding vehicles 20 and 30 having the tail lamps 21 and 31 and the reflectors 22 and 32.

  The operation flow in the vehicle detection device 3 may be the same as the operation flow in the first embodiment, or may be the same as the operation flow in the second embodiment.

  The vehicle detection device 3 has the same effects as the vehicle detection device 1 according to the first embodiment or the vehicle detection device 2 according to the second embodiment.

  The present invention is not limited to the embodiment described above. For example, the vehicle detection devices 1, 2, and 3 can detect not only four wheels but also two-wheel preceding vehicles. In this case, the preceding vehicle candidate area extraction units 14a and 14b detect a pixel set as a tail lamp candidate from the shape of the tail lamp of the two-wheeled vehicle for a set area of pixels having a luminance of 1 in the binarized image, A surrounding rectangular area including the tail lamp candidate is extracted as a preceding vehicle candidate area T. Thereafter, a two-wheel preceding vehicle having a tail lamp and a reflector is detected from the difference between the luminance at each pixel of the visible light image A and the luminance at each pixel of the near-infrared image B.

  In addition, in obtaining the difference between the luminance in each pixel of the visible light image A and the luminance in each pixel B of the near-infrared image, an aspect may be adopted in which the difference is obtained after binarizing each image in advance. In this case, the calculation processing load in the vehicle detection ECUs 10, 11, and 12 can be reduced.

  The preceding vehicle candidate area T is not limited to a rectangular area, and may be an elliptical or oval area, and the shape of the area is not particularly limited.

It is a lineblock diagram showing the vehicle detection device concerning a 1st embodiment. It is a graph which shows the wavelength range which can be imaged with the CCD camera shown in FIG. It is a figure which shows an example of the visible light image in front of the own vehicle. It is a figure which shows an example of the near-infrared image in front of the own vehicle. It is a figure which shows the difference in the preceding vehicle candidate area | region of the visible light image of FIG. 3, and the near-infrared image of FIG. It is a figure which shows the other example of the visible light image in front of the own vehicle. It is a figure which shows the other example of the near-infrared image in front of the own vehicle. It is a figure which shows the difference in the preceding vehicle candidate area | region of the visible light image of FIG. 6, and the near-infrared image of FIG. It is a flowchart which shows the flow of operation | movement of the vehicle detection apparatus which concerns on 1st Embodiment. It is a flowchart which shows the flow of operation | movement of the vehicle detection apparatus which concerns on 2nd Embodiment. It is a block diagram which shows the vehicle detection apparatus which concerns on 3rd Embodiment. (A) is a graph which shows the wavelength area | region which can be imaged with the visible light camera shown in FIG. (B) is a graph which shows the wavelength range which can be imaged with the near-infrared camera shown in FIG. It is a figure which shows the difference in the preceding vehicle candidate area | region of the visible light image of FIG. 3, and the near-infrared image of FIG. 4 which concern on 3rd Embodiment. It is a figure which shows the difference in the preceding vehicle candidate area | region of the visible light image of FIG. 6, and the near-infrared image of FIG. 7 which concern on 3rd Embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1, 2, 3 ... Vehicle detection apparatus, 4 ... CCD camera (visible light image acquisition means, near-infrared image acquisition means), 5 ... Near-infrared light projector (infrared light projection means), 6 ... Driving support device, 7 ... Visible light Camera (visible light image acquisition means), 8 ... Near infrared camera (infrared image acquisition means), 10, 11, 12 ... Vehicle detection ECU, 13a, 13b ... Visible light image acquisition unit (visible light image acquisition means), 14a, 14b ... preceding vehicle candidate area extraction unit (extraction means), 15a, 15b ... near infrared image acquisition unit (infrared image acquisition means), 16a, 16b, 17 ... preceding vehicle detection unit (preceding vehicle detection means), 20, 30 ... A preceding vehicle, 21, 31 ... tail lamp, 22, 32 ... reflector, illumination ... 26, road reflector ... 28.

Claims (6)

A vehicle detection device for detecting a preceding vehicle traveling in front of a vehicle,
Visible light image acquisition means for acquiring a visible light image in front of the vehicle;
Infrared light projecting means for irradiating infrared light in front of the vehicle;
Infrared image acquisition means for acquiring an infrared image in front of the vehicle irradiated with infrared rays by the infrared projection means;
Vehicle detection comprising: preceding vehicle detection means for detecting a preceding vehicle based on the infrared image acquired by the infrared image acquisition means and the visible light image acquired by the visible light image acquisition means apparatus.   In the preceding vehicle detection means, a difference in luminance value between pixels in the infrared image acquired by the infrared image acquisition means and the visible light image acquired by the visible light image acquisition means is greater than or equal to a first threshold value. The vehicle detection device according to claim 1, wherein the preceding vehicle is detected based on the vehicle. An extraction means for extracting a preceding vehicle candidate area based on at least one of the visible light image acquired by the visible light image acquisition means and the infrared image acquired by the infrared image acquisition means;
The vehicle detection device according to claim 1, wherein the preceding vehicle detection unit detects the preceding vehicle based on a preceding vehicle candidate area extracted by the extraction unit.   The infrared light projecting unit switches from a light-off state to a light-up state when the extraction unit extracts the preceding vehicle candidate area based on the visible light image acquired by the visible light image acquisition unit. The vehicle detection device according to claim 3.   The infrared light projecting unit switches from a lighting state to a non-lighting state when the extraction unit extracts the preceding vehicle candidate region based on the infrared image acquired by the infrared image acquisition unit. 4. The vehicle detection device according to 3.   The preceding vehicle detection unit is configured such that, after the infrared light projecting unit is switched to a lighting state, a luminance change amount per predetermined time in the infrared image acquired by the infrared image acquisition unit is equal to or less than a second threshold value. The vehicle detection apparatus according to claim 1, wherein the preceding vehicle is detected based on the infrared image.

Images