JP2008174028A - Target detection system and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a target detection system and a target detection method capable of preventing erroneous identification of a target with high accuracy by verifying an identification result obtained by using a far infrared imaging device again by means of a near infrared imaging device (visible light imaging device). <P>SOLUTION: An obstacle candidate area is specified based on image data obtained from the far infrared imaging device 1. When it is determined that the specified area is a target area corresponding to a target, positional coordinates, a feature quantity, and vehicle positional information at the acquisition time of the image data from which the target area is specified are recorded. When a predetermined condition is satisfied, the far infrared imaging device 1 for acquiring image data is switched to the near infrared imaging device 2, and image data at different times are acquired. The obstacle candidate area is identified again, and if it disagrees with the positional coordinates of the stored target area, the stored positional coordinates of the target area, its feature quantity, and the vehicle positional information at the acquisition time of the image data from the target area is specified are deleted. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an object detection system and an object detection method that can prevent erroneous recognition of an object to be recognized based on image data captured by an imaging device that captures the periphery of a vehicle.

  Many object detection systems that detect an object such as a pedestrian that may collide with a traveling vehicle and output a warning to the driver have been developed. A conventional object detection system extracts an object such as a pedestrian who has a high possibility of colliding with a vehicle from an image around the vehicle captured by an imaging device such as an infrared camera, and the object exists. Information to the driver.

  For example, in Patent Document 1, the distance to a target object such as a pedestrian is calculated by obtaining the parallax of the target object in an image around the host vehicle captured by a pair of left and right stereo cameras. Based on the calculated distance, position, moving direction, etc., an object that is highly likely to collide is detected and an alarm is output.

However, if there are artificial structures such as roadside utility poles, streetlight poles, vending machines, etc. that show a temperature distribution similar to that of objects such as pedestrians, it is easy to misrecognize the artificial structures as objects Therefore, for example, a warning is output to the driver even though there is no pedestrian, and there is a problem that the driver may hinder driving. In order to solve such a problem, for example, in Patent Document 2, the feature and position of the detected object can be stored in the storage means by the operation of the driver, and the stored object is detected again. In such a case, there is disclosed an object detection device that stops outputting an alarm for the object. In Patent Document 2, it is possible for the driver to consciously select an object to be excluded from the alarm output target.
JP 2001-006096 A JP 2004-106682 A

  However, in the above-described object detection device, it is necessary for the driver to select an object that is misrecognized as a pedestrian, and it is extremely dangerous for the driver to perform such an operation while traveling. There was a problem.

  The present invention has been made in view of such circumstances, and avoids erroneous recognition of an object with high accuracy by automatically re-verifying recognition results obtained using an imaging device at different points in time. It is an object of the present invention to provide an object detection system and an object detection method that can be operated and do not require an operation by a driver.

  In order to achieve the above object, an object detection system according to a first aspect of the present invention includes an imaging device that images the periphery of a vehicle, and a detection that acquires an image captured by the imaging device and detects an object included in the image. An object detection system comprising a device, comprising: a vehicle position detection device that detects vehicle position information, wherein the detection device identifies an obstacle candidate region based on one image data acquired from the imaging device An object candidate area specifying means; a means for determining whether the object area corresponds to the object based on a feature amount of one image data corresponding to the specified obstacle candidate area; and If it is determined to be an object area, the position coordinates of the object area and means for recording the position information of the vehicle at the time when one image data specifying the object area is acquired; and the one image data Acquired Means for acquiring other image data at a position corresponding to the recorded vehicle position information at a different time from the time, and means for respecifying an obstacle candidate area based on the other image data; When it is determined that the position coordinates of the re-specified obstacle candidate area and the position coordinates of the recorded object area are substantially matched, and the means do not substantially match, Means for deleting the recorded position coordinates of the object area and the position information of the vehicle at the time when one image data specifying the object area is acquired, and the recorded position coordinates of the object area And, based on the position information of the vehicle, misrecognition of the object is prevented.

  An object detection system according to a second aspect of the present invention includes a first imaging device that images the periphery of a vehicle, and a detection device that acquires an image captured by the first imaging device and detects an object included in the image. The object detection system includes a vehicle position detection device that detects vehicle position information, and a second imaging device that has a wavelength band that can be imaged different from that of the first imaging device, and the detection device includes the first imaging device. Corresponding to an object based on an obstacle candidate area specifying means for specifying an obstacle candidate area based on one image data acquired from the apparatus, and a feature amount of the one image data corresponding to the specified obstacle candidate area A means for determining whether or not the object area is to be acquired, and if the means determines that the object area is, the position coordinates of the object area and one image data specifying the object area are acquired. The position of the vehicle at the time Means for recording information, and if the predetermined condition is satisfied, selects the second imaging device; if not, the means for selecting the first imaging device is selected. In the first imaging device or the second imaging device, another image at a time different from the time when the first imaging device acquired one image data at a position corresponding to the recorded vehicle position information. Means for acquiring data, means for respecifying an obstacle candidate area based on the other image data, position coordinates of the respecified obstacle candidate area, and position coordinates of the recorded object area Means for determining whether or not they substantially match, and when the means determines that they do not substantially match, the recorded position coordinates of the object area and one image data specifying the object area Delete vehicle location information at the time of obtaining That a means, based on the position coordinates and the position information of the vehicle of the recorded object region, and wherein the you have to prevent the erroneous recognition of the object.

  In the object detection system according to a third aspect of the invention, the first imaging device is a far-infrared imaging device, the second imaging device is a visible light imaging device or a near-infrared imaging device, and an illuminance sensor that detects ambient illuminance is provided. A criterion for determining whether or not the predetermined condition is satisfied is whether or not the detection value of the illuminance sensor exceeds a predetermined value, and it is determined that the detection value of the illuminance sensor exceeds the predetermined value If the second imaging device is selected, the first imaging device is selected when it is determined that the second imaging device is equal to or less than a predetermined value.

  In the object detection system according to a fourth aspect of the invention, the first imaging device is a far-infrared imaging device, the second imaging device is a visible light imaging device or a near-infrared imaging device, and a switch for turning on and off a vehicle headlight is provided. A criterion for determining whether or not the predetermined condition is satisfied is whether or not the switch is in an off state. When the switch is determined to be in an off state, the second imaging device is When it is selected and it is determined that the switch is on, the first imaging device is selected.

  In the object detection system according to a fifth aspect of the present invention, the first imaging device is a far-infrared imaging device, and the second imaging device is a visible light imaging device or a near-infrared imaging device, and has the predetermined condition The determination criterion is whether or not the average value of the luminance value of the image captured by the second imaging device exceeds a predetermined value, and it is determined that the average value of the luminance value exceeds the predetermined value. The second imaging device is selected, and when it is determined that the average value is equal to or less than a predetermined value, the first imaging device (or the second imaging device) is selected. To do.

  An object detection method according to a sixth aspect of the present invention is an object detection method for acquiring an image captured by an image capturing apparatus that captures the periphery of a vehicle and detecting an object included in the image. The obstacle candidate area is identified based on the image data of the image, and it is determined whether the object area corresponds to the object based on the feature amount of one image data corresponding to the identified obstacle candidate area. When it is determined that the target area is the object area, the position coordinates of the target area and the position information of the vehicle at the time when one image data specifying the target area is acquired are recorded. Acquire other image data at a position corresponding to the recorded vehicle position information at a different time from the acquired time, re-specify the obstacle candidate area based on the other image data, and re-specify Of the obstacle candidate area It is determined whether or not the position coordinates of the target area and the recorded object area are substantially the same, and when it is determined that they are not approximately the same, the position coordinates of the object area that is recorded and the target The vehicle position information at the time when one image data specifying the object area is acquired is deleted.

  In the first invention and the sixth invention, an image captured by an imaging device such as a visible light imaging device, a far-infrared imaging device, or a near-infrared imaging device that captures the periphery of the vehicle is acquired, and an object included in the image is obtained. To detect. More specifically, the obstacle candidate area is specified based on one image data acquired from the imaging apparatus. When it is determined whether the target area corresponds to the target object based on the feature amount of one image data corresponding to the identified obstacle candidate area, The position coordinates of the area and the position information of the vehicle at the time when one image data specifying the object area is acquired are recorded. Next, another image data is acquired at a position corresponding to the recorded vehicle position information at a time different from the time when the one image data is acquired, and the obstacle candidate area is based on the other image data. Respecify. It is determined whether or not the position coordinates of the re-specified obstacle candidate area and the position coordinates of the recorded object area are substantially matched. The position coordinates of the object area and the position information of the vehicle at the time when one image data specifying the object area is acquired are deleted. When an obstacle candidate area is identified again at a different timing at the same position as the object area once recognized as an object, there is a stationary object that is easily misrecognized, such as a vending machine. By judging that the possibility is high and keeping a record of the object, it is possible to avoid erroneously recognizing that the object is a pedestrian or the like in the object detection process using the imaging device from the next time onward. it can. In addition, it is not necessary for the driver to register information consciously, and it is possible to continue driving safely.

  In the second invention, the first imaging device and the second imaging device having different wavelength bands that can be imaged are selectively used depending on the situation. Since an image picked up using either the first image pickup device or the second image pickup device selected according to the situation is used, the detection accuracy of the object is improved. The first imaging device and the second imaging device are, for example, a visible light imaging device, a far infrared imaging device, a near infrared imaging device, or the like.

  In the third invention, when the detection value of the illuminance sensor that detects the ambient illuminance exceeds a predetermined value, the second imaging device, which is a visible light imaging device or a near-infrared imaging device, is selected and is equal to or less than the predetermined value. The first imaging device that is a far-infrared imaging device is selected. Thereby, when the brightness around the vehicle is a predetermined value or more, for example, in the daytime, the obstacle candidate area can be specified by using the second imaging device, and the recognition result of the object area is verified. can do. Further, when the brightness around the vehicle is equal to or lower than a predetermined value, for example, at night, by using the first imaging device, it is possible to reliably detect the presence of an object that cannot be confirmed by the second imaging device. And the recognition result of the object area can be verified.

  In the fourth invention, when the switch for turning on / off the headlight of the vehicle is in the off state, the second imaging device which is the visible light imaging device or the near infrared imaging device is selected, and when the switch is in the on state, the far infrared imaging The first imaging device, which is a device, is selected. Thereby, when the vehicle headlight is not necessary, for example, in the daytime, the obstacle candidate area can be specified by using the second imaging device, and the recognition result of the object area can be verified. . In addition, when a vehicle headlight is necessary, for example, at night, by using the first imaging device, it is possible to reliably detect the presence of an object that cannot be confirmed by the second imaging device, The recognition result of the object area can be verified.

  In the fifth aspect of the invention, when it is determined that the average value of the luminance values of the images picked up by the second image pickup device which is a visible light image pickup device or a near infrared image pickup device is equal to or less than a predetermined value, When the one-line imaging device is selected and it is determined that the average value is exceeded, the second imaging device is selected. Thus, when the average value of the luminance values of the images captured by the second imaging device is not more than a predetermined value, that is, when the brightness of the surroundings of the vehicle is not more than a certain value, an obstacle can be obtained by using the first imaging device. Candidate areas can be specified, and recognition results during night driving can be verified. In addition, when the average value of the luminance values of the images captured by the second imaging device exceeds a predetermined value, that is, when the brightness around the vehicle is a certain value or more, the first imaging device is used. The presence of an object that cannot be confirmed by the second imaging apparatus can be reliably detected, and the recognition result of the object area can be verified.

  According to the present invention, when an obstacle candidate area is specified again at a different timing at the same position as the object area once recognized as an object such as a pedestrian, a vending machine or the like is located at the position. It is determined that there is a high possibility that there is a stationary object that is likely to be erroneously recognized, and the position coordinates of the object area and the position coordinates of the vehicle that acquired the image data specifying the object area are recorded. Thus, it is possible to avoid erroneous recognition that the object is a pedestrian or the like in the object detection process using the imaging device from the next time onward. In addition, it is not necessary for the driver to register information consciously, and it is possible to continue driving safely.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In the following embodiment, an obstacle such as a pedestrian or a bicycle is detected in front of the vehicle based on an image captured by the far-infrared imaging device (imaging device) while the vehicle is traveling. A case will be described as an example.

(Embodiment 1)
FIG. 1 is a schematic diagram showing a configuration of an object detection system according to Embodiment 1 of the present invention. In the first embodiment, a set of far-infrared imaging devices 1 and 1 that capture an image of the periphery of a vehicle during traveling are mounted in a front grille (also possible in a bumper) near the center in front of the vehicle. The far-infrared imaging devices 1 and 1 are imaging devices using infrared light having a wavelength of 7 to 14 micrometers.

  In FIG. 1, far-infrared imaging devices 1 and 1 are juxtaposed in a substantially symmetrical position in the left-right direction within a front grill of a vehicle. Image data captured by the far-infrared imaging devices 1 and 1 is transmitted to a detection device 3 connected via a video cable 7 corresponding to an analog video system such as NTSC or a digital video system.

  A display device 4 having an operation unit is connected to the detection device 3 via a cable 8 corresponding to a video system such as NTSC, VGA, DVI, etc., and an audible warning is given by sound, sound effect, etc. An output device such as an alarm device 5 is connected via an in-vehicle LAN cable 6 compliant with CAN.

  In addition, a vehicle position detecting device 9 capable of performing GPS communication for detecting the position of the vehicle is provided on the ceiling portion of the vehicle. The vehicle position detection device 9 is connected to the detection device 3 via the in-vehicle LAN cable 6 conforming to CAN, and transmits the latitude information and longitude information of the vehicle in time series.

  FIG. 2 is a block diagram illustrating a configuration of the far-infrared imaging device 1 of the object detection system according to Embodiment 1 of the present invention. The image pickup unit 11 includes image pickup elements that convert optical signals into electric signals in a matrix. As an imaging device for infrared light, an infrared sensor such as a vanadium oxide bolometer type using a micromachining technique or a BST (Barium-Strontium-Titanium) pyroelectric type sensor is used. The image capturing unit 11 reads an infrared light image around the vehicle as a luminance signal, and transmits the read luminance signal to the signal processing unit 12.

  The signal processing unit 12 is an LSI, converts the luminance signal received from the image imaging unit 11 into a digital signal such as an RGB signal, and performs processing for correcting variations in the imaging device, correction processing for a defective device, gain control processing, and the like. And stored in the image memory 13 as image data. Needless to say, it is not essential to temporarily store the image data in the image memory 13, and the image data may be transmitted directly to the detection device 3 via the video output unit 14.

  The video output unit 14 is an LSI, and outputs image data to the detection device 3 via a video cable 7 compatible with an analog video system such as NTSC or a digital video system.

  FIG. 3 is a block diagram showing the configuration of the detection device 3 of the object detection system according to Embodiment 1 of the present invention. The detection device 3 is an image processing ECU, and includes at least a video input unit 31a, a video output unit 31b, a communication interface unit 31c, an image memory 32, and an LSI 33 incorporating a RAM 331.

  The video input unit 31 a inputs video signals from the far infrared imaging devices 1 and 1. The video input unit 31a stores the image data input from the far-infrared imaging devices 1 and 1 in the image memory 32 in synchronization with each frame. The video output unit 31 b outputs image data to the display device 4 such as a liquid crystal display via the video cable 8.

  The communication interface unit 31 c transmits an output signal such as a synthesized sound to the alarm device 5 such as a buzzer or a speaker via the in-vehicle LAN cable 6. Further, the communication interface unit 31 c receives latitude information and longitude information indicating the position of the vehicle from the vehicle position detection device 9 via the in-vehicle LAN cable 6 at regular time intervals, and stores them in the RAM 331.

  The image memory 32 is an SRAM, flash memory, SDRAM or the like, and stores image data input from the far-infrared imaging devices 1 and 1 via the video input unit 31a.

  The LSI 33 that performs image processing reads the image data stored in the image memory 32 in units of frames and specifies an obstacle candidate area. The LSI 33 estimates a representative point of the identified obstacle candidate area, for example, the position coordinates of the center of gravity, based on the parallax of the read image data. The position coordinates of the representative point of the obstacle candidate area are not limited to the method of calculating and estimating the distance based on the parallax by the imaging devices installed on the left and right. A method of directly obtaining a sonic device and the like may be used. Needless to say, when a radar device, an ultrasonic device, or the like is provided, it is sufficient to install one far-infrared imaging device 1.

  The LSI 33 extracts the feature amount of the obstacle candidate area based on the read image data, and determines whether or not an object such as a pedestrian exists based on the extracted feature amount. When the LSI 33 determines that an object such as a pedestrian is present, the LSI 33 is a basis for the position coordinates, the feature amount, and the calculation of the representative point of the object area recognized as the object such as the pedestrian. The position coordinates of the vehicle at the time of capturing the image data are recorded in the RAM 331.

  The LSI 33 reads out image data stored in the image memory 32 captured at the same vehicle position in different time zones in units of frames, and specifies an obstacle candidate region again. For example, when the image data that is the basis of the information on the recorded object area is captured at night, the image data captured in the daytime is read out. The LSI 33 estimates the position coordinates of the identified obstacle candidate area based on the parallax of the read image data, and queries the RAM 331 to determine whether the position coordinates of the object area recorded are different. to decide.

  When the LSI 33 determines that the position coordinates of the two areas are different, it determines that the object detection result is appropriate, that is, the information regarding the object area recorded in the RAM 331 is information regarding the pedestrian. If it is determined that the position coordinates are substantially the same, the LSI 33 determines that the recognition that the target object such as a pedestrian has been detected is an error, and the information on the target area recorded in the RAM 331 includes information such as the artificial structure. Judged as information on a stationary object.

  Hereinafter, detailed processing of the LSI 33 of the detection device 3 of the object detection system according to the first embodiment of the present invention will be described. 4 and 5 are flowcharts showing the procedure of the object detection information recording process of the LSI 33 of the detection apparatus 3 of the object detection system according to the first embodiment of the present invention.

  The LSI 33 acquires vehicle position coordinates (latitude information and longitude information) from the vehicle position detection device 9 (step S401). Next, the LSI 33 acquires image data captured by the far-infrared imaging devices 1 and 1 (step S402).

  Then, the LSI 33 divides the image data captured by the one far-infrared imaging device 1 acquired from the image memory 32 into small regions of a predetermined size in order to specify an obstacle candidate region (step S403). The method for setting the small area to be divided is not particularly limited. For example, you may set as a square area | region which consists of m * n (m and n are natural numbers) pixels. In the first embodiment, a small area is set as a square area composed of 3 × 3 pixels.

  The LSI 33 uses a divided small area as a reference area, an area corresponding to one reference area in image data captured by one far-infrared imaging device 1, that is, an area indicating the same object (hereinafter referred to as “corresponding area”). Is searched from the image data captured by the other far-infrared imaging device 1. Specifically, one reference area in the image data captured by one far-infrared imaging device 1 and an area of the same size as the reference area in the image data captured by the other far-infrared imaging device 1 (hereinafter referred to as “search”). (Referred to as “region”) (step S404), and it is determined whether or not the correlation value is greater than a predetermined value (step S405).

  When the LSI 33 determines that the correlation value between the reference area and the search area is smaller than the predetermined value (step S405: NO), the LSI 33 changes the search area, returns the process to step S404, and repeats the above-described process. When the LSI 33 determines that the correlation value is greater than the predetermined value (step S405: YES), the LSI 33 regards the search area as the corresponding area and calculates the parallax between the reference area and the corresponding area (step S406). Specifically, a reference region in an image captured by one far-infrared imaging device 1, for example, a far-infrared imaging device 1 installed on the right, and a corresponding region of the reference region, that is, the other far-infrared imaging device 1, For example, the parallax with the search area having a correlation value larger than a predetermined value in the image captured by the far-infrared imaging device 1 installed on the left is calculated (step S406).

  The LSI 33 determines whether or not parallax has been calculated for all reference regions in the image captured by one far-infrared imaging device 1 (step S407), and the LSI 33 determines that there is an uncalculated reference region. In the case (step S407: NO), the LSI 33 changes the reference area (step S408), returns the process to step S404, and repeats the above-described process. When the LSI 33 determines that the parallax has been calculated for all the reference areas in the image (step S407: YES), the LSI 33 combines adjacent small areas having substantially the same parallax and identifies it as an obstacle candidate area ( Step S409).

  The LSI 33 extracts a feature amount used to determine whether or not an object such as a pedestrian exists in the specified obstacle candidate area, and the obstacle candidate area is determined to be a pedestrian or the like in accordance with a determination rule. It is determined whether or not the target area indicates the target object. As an example, a correlation value obtained by pattern matching between pixel data of an obstacle candidate area and pixel data of a template indicating a standard human temperature distribution is calculated as a feature amount (step S410), and a predetermined value set as a determination rule It is determined whether or not the correlation value calculated from the value is large (step S411). When the LSI 33 determines that the correlation value is equal to or less than the predetermined value (step S411: NO), the LSI 33 determines that the area is not an object area indicating the presence of an object such as a pedestrian, and the process returns to step S410. The same process is repeated for the next obstacle candidate area.

  When the LSI 33 determines that the correlation value is greater than the predetermined value (step S411: YES), the LSI 33 determines that the obstacle candidate area is an object area indicating an object such as a pedestrian, and calculates The position coordinates of the representative point of the object area, the extracted feature amount, and the position coordinates of the vehicle when the image data specifying the object area are acquired are recorded in the RAM 331 (step S412).

  It should be noted that as the feature quantity used by the LSI 33 to determine whether or not an object such as a pedestrian exists, pixel data of a template indicating a standard human temperature distribution and pixel data of an obstacle candidate area When the correlation value R calculated by matching is used, the correlation value R is calculated based on (Equation 1).

  In (Expression 1), N is the total number of pixels of the template, k is an integer of 0 ≦ k ≦ (N−1), Fk is the pixel value of the kth pixel in the template, and Gk is the obstacle candidate area. The pixel values of the kth pixel in FIG.

  Note that the feature amount used to determine whether or not the identified obstacle candidate region is an image that should be recognized as an object such as a pedestrian is not limited to the correlation value described above. It is possible to specify the area that is recognized as an object such as a pedestrian, such as the size, aspect ratio, average pixel value, and variance of the area that is recognized as an object such as a pedestrian. Any feature amount may be used, or a combination of these may be determined. Further, the determination rule is not limited to threshold processing, and determination may be performed by a learning machine such as a support vector machine or boosting.

  When the LSI 33 determines that the obstacle candidate area is an object area indicating the presence of an object such as a pedestrian, the LSI 33 displays a signal indicating that the object such as a pedestrian has been detected on the display device 4 or Output to the alarm device 5. The display device 4 outputs a warning display such as outputting a display indicating that an object such as a pedestrian has been detected or blinking the warning display. The alarm device 5 notifies the driver of the presence of an object such as a pedestrian by sounding a buzzer or the like.

  6 and 7 are flowcharts showing a procedure of erroneous recognition avoidance processing of the LSI 33 of the detection apparatus 3 of the object detection system according to Embodiment 1 of the present invention. The LSI 33 acquires the vehicle position coordinates from the vehicle position detection device 9 (step S601), and determines whether or not the vehicle position coordinates are substantially coincident with the vehicle position coordinates recorded in the RAM 331 (step S602). When it is determined that the LSI 33 is different (step S602: NO), the LSI 33 returns to step S601 and acquires the position coordinates of the vehicle. When it is determined that the LSIs 33 substantially match (step S602: YES), the LSI 33 acquires image data captured by the far-infrared imaging devices 1 and 1 (step S603).

  When the vehicle travels at the same position at different times, there is almost no obstacle candidate region having the same feature amount at the position where the presence of an object such as a pedestrian is detected last time. This is because pedestrians move. However, when detecting artificial structures such as roadside utility poles, streetlight poles, vending machines, etc. that show a temperature distribution similar to that of pedestrians, etc., even if the times taken are different, they are substantially the same. An object having substantially the same feature amount at the position of is detected.

  Therefore, when the vehicle travels at the same position at different times, and there is an obstacle candidate area having the same feature amount at the position where the presence of the object such as the previous pedestrian is detected, the object such as the previous pedestrian Judging that the presence of an object itself is an error and leaving a record of the position and feature amount of the previous object area, it will be erroneously recognized as an object such as a pedestrian after the next time. You can avoid that. Conversely, when the vehicle travels at the same position at different times and there is no obstacle candidate area having the same feature amount at the position detected as the object of the previous pedestrian or the like, the object of the previous pedestrian or the like It is determined that the presence of the object is detected correctly, and the position and feature amount of the object area recorded in the RAM 331 are deleted.

  In order to implement the above-described processing, the LSI 33 divides the image data acquired from the image memory 32 into small areas having a predetermined size (step S604). The method for setting the small area to be divided is not particularly limited. For example, you may set as a square area | region which consists of m * n (m and n are natural numbers) pixels. In the first embodiment, a small area is set as a square area composed of 3 × 3 pixels.

  The LSI 33 uses a divided small area as a reference area, a reference area in image data captured by one far-infrared imaging device 1, and a search area in image data captured by the other far-infrared imaging apparatus 1. A normal cross-correlation value is calculated (step S605), and it is determined whether or not there is a search region having a correlation value greater than a predetermined value (step S606).

  When the LSI 33 determines that there is no search area larger than the predetermined value (step S606: NO), the LSI 33 changes the search area, returns the process to step S605, and repeats the above-described process. When the LSI 33 determines that there is a search area larger than the predetermined value (step S606: YES), the LSI 33 has a parallax between the reference area and the search area, that is, one of the far-infrared imaging devices 1, for example, a distant place installed on the right. A correlation value is predetermined in an image captured by a reference region in an image captured by the infrared imaging device 1 and the other far infrared imaging device 1 corresponding to the reference region, for example, a far infrared imaging device 1 installed on the left. The parallax with the search area larger than the value is calculated (step S607).

  The LSI 33 determines whether or not the parallax has been calculated for all the small areas in the image (step S608), and if the LSI 33 determines that there is a small area that has not been calculated (step S608: NO), the LSI 33 Then, the small area is changed (step S609), the process returns to step S605, and the above-described process is repeated. When the LSI 33 determines that the parallax has been calculated for all the small regions in the image (step S608: YES), the LSI 33 combines adjacent small regions whose parallax is substantially the same and identifies it as an obstacle candidate region ( Step S610).

  Next, the LSI 33 estimates the position coordinates of the center of gravity position (representative point) of the obstacle candidate area by estimating the distance to the obstacle candidate area based on the parallax images of the left and right imaging devices (step S611). . Of course, the distance to the obstacle that actually exists and the position coordinates of the representative point of the obstacle may be directly obtained by a laser device, an ultrasonic device, or the like.

  Next, the LSI 33 determines whether or not the estimated position coordinates of the obstacle candidate area are different from the position coordinates of the object area recorded in the RAM 331 (step S612). If it is determined that the LSI 33 is different (step S612: YES), the LSI 33 determines that the process for detecting the presence of the object such as the pedestrian is the correct process, and the object area recorded in the RAM 331 is determined. The position coordinates, the feature amount of the image data corresponding to the object area, and the position coordinates of the vehicle when the image data specifying the object area are acquired are deleted (step S613). Whether or not the position coordinates are different is determined based on whether or not the estimated position coordinates of the obstacle candidate area are included within a predetermined range from the position coordinates of the object area recorded in the RAM 331.

  When it is determined that the LSI 33 substantially matches (step S612: NO), the LSI 33 determines that the process of detecting the presence of an object such as a pedestrian is incorrect last time and that the detected object is an artificial structure. The record of the object area in the RAM 331 is left (step S614). From the next time, the position coordinates of the object area recorded in the RAM 331, the feature amount of the image data corresponding to the object area, and the position coordinates of the vehicle at the time of obtaining the image data specifying the object area are collated. By doing so, it becomes possible to avoid erroneously recognizing an object that is not an object such as a pedestrian but an artificial structure as an object such as a pedestrian. Therefore, the LSI 33 does not output a signal indicating that the presence of an object such as a pedestrian has been detected to the display device 4 or the alarm device 5, and the driver is bothered by unnecessary warning display, alarm sound, and the like. You can drive safely without any problems.

  As described above, according to the first embodiment, when an object area that has been recognized as an object such as a pedestrian is not an object such as a pedestrian at a separate timing, The position coordinates of the vehicle that has determined that the recognition itself as an object such as a pedestrian is an error, and acquired the position coordinates, the feature amount of the object area, and the image data specifying the object area Can be prevented from being erroneously recognized as an object such as a pedestrian in the subsequent object detection process. In addition, it is not necessary for the driver to register information consciously, and it is possible to continue driving safely.

  In the first embodiment described above, the LSI 33 of the detection device 3 executes the above-described processing, but a control device may be provided separately, or a control device of another device may also be used.

Further, as described above, the determination is not limited to whether or not the position coordinates of the representative point of the obstacle candidate area and the position coordinates of the representative point of the object area substantially coincide with each other. Needless to say, the determination of whether or not the feature amount of the region and the feature amount of the object region substantially coincide may be used together.
Furthermore, in the first embodiment, the case where the far-infrared imaging device is provided has been described, but other imaging devices such as a visible light imaging device and a near-infrared imaging device having a wavelength band that can be imaged are different from those of the far-infrared imaging device. May be. However, it is more preferable to provide the infrared imaging device because there is a low possibility of erroneously recognizing a stationary object other than a pedestrian or the like that emits infrared rays.

(Embodiment 2)
FIG. 8 is a schematic diagram showing a configuration of an object detection system according to Embodiment 2 of the present invention. In the second embodiment, a set of far-infrared imaging devices 1, 1 (first imaging device) and near-infrared imaging devices 2, 2 (second imaging device) that capture an image of the periphery of the vehicle while traveling are used for the vehicle. It is mounted in the front grille near the front center (possible in a bumper). The far infrared imaging devices 1 and 1 are imaging devices using infrared light having a wavelength of 7 to 14 micrometers, and the near infrared imaging devices 2 and 2 are red having a wavelength of 0.8 to 3 micrometers. This is an imaging device using external light. Further, instead of the near infrared imaging devices 2 and 2, a visible light imaging device using visible light having a wavelength of 0.4 to 0.8 micrometers may be provided.

  In FIG. 8, the far-infrared imaging devices 1 and 1 and the near-infrared imaging devices 2 and 2 are juxtaposed at substantially symmetrical positions in the left-right direction in the front grill of the vehicle. Image data captured by the far-infrared imaging devices 1 and 1 and the near-infrared imaging devices 2 and 2 are connected via a video cable 7 corresponding to an analog video system such as NTSC or a digital video system. Sent to.

  In addition to the far-infrared imaging devices 1, 1, the near-infrared imaging devices 2, and 2, the detection device 3 includes a display device 4 having an operation unit via a cable 8 that supports video systems such as NTSC, VGA, and DVI. An output device such as an alarm device 5 that issues an audible warning by voice, sound effect, or the like is connected via an in-vehicle LAN cable 6 that conforms to CAN.

  In addition, a vehicle position detecting device 9 capable of performing GPS communication for detecting the position of the vehicle is provided on the ceiling portion of the vehicle. The vehicle position detection device 9 is connected to the detection device 3 via the in-vehicle LAN cable 6 conforming to CAN, and transmits the latitude information and longitude information of the vehicle in time series.

  The image pickup unit 11 includes image pickup elements that convert optical signals into electric signals in a matrix. As an imaging device for infrared light, an infrared sensor such as a vanadium oxide bolometer type or a BST pyroelectric type using a micromachining technique is used. The image capturing unit 11 reads an infrared light image around the vehicle as a luminance signal, and transmits the read luminance signal to the signal processing unit 12.

  The signal processing unit 12 is an LSI, converts the luminance signal received from the image imaging unit 11 into a digital signal such as an RGB signal, and performs processing for correcting variations in the imaging device, correction processing for a defective device, gain control processing, and the like. And stored in the image memory 13 as image data. Needless to say, it is not essential to temporarily store the image data in the image memory 13, and the image data may be transmitted directly to the detection device 3 via the video output unit 14.

  The video output unit 14 is an LSI, and outputs image data to the detection device 3 via a video cable 7 compatible with an analog video system such as NTSC or a digital video system.

  The configuration of the near-infrared imaging device 2 is the same as that of the far-infrared imaging device 1, and the wavelength of infrared light that can be detected by an infrared sensor such as a CCD or CMOS used as an imaging element is in the vicinity of 0.8 to 3 micrometers. The only difference is that it is infrared light.

  FIG. 9 is a block diagram showing the configuration of the detection device 3 of the object detection system according to Embodiment 2 of the present invention. The detection device 3 is an image processing ECU, and includes at least a video input unit 31a, a video output unit 31b, a communication interface unit 31c, an image memory 32, and an LSI 33 incorporating a RAM 331.

  The video input unit 31 a inputs video signals from the far infrared imaging devices 1 and 1 and the near infrared imaging devices 2 and 2. The video input unit 31a stores the image data input from the far-infrared imaging devices 1 and 1 and the near-infrared imaging devices 2 and 2 in the image memory 32 in synchronization with each frame. The video output unit 31 b outputs image data to the display device 4 such as a liquid crystal display via the video cable 8.

  The communication interface unit 31 c transmits an output signal such as a synthesized sound to the alarm device 5 such as a buzzer or a speaker via the in-vehicle LAN cable 6. Further, the communication interface unit 31 c receives latitude information and longitude information indicating the position of the vehicle from the vehicle position detection device 9 via the in-vehicle LAN cable 6 at regular time intervals, and stores them in the RAM 331.

  The image memory 32 is an SRAM, a flash memory, an SDRAM, or the like, and stores image data input from the far infrared imaging devices 1 and 1 and the near infrared imaging devices 2 and 2 via the video input unit 31a.

  The LSI 33 that performs image processing reads the image data stored in the image memory 32 in units of frames and specifies an obstacle candidate area. The LSI 33 reads the image data acquired from the far-infrared imaging devices 1 and 1 based on a predetermined condition, for example, whether or not the ambient brightness is a certain level or more, or the image data from the near-infrared imaging devices 2 and 2. Is switched, and based on the parallax of the read image data, the representative point of the specified obstacle candidate area, for example, the position coordinates of the center of gravity is estimated. The position coordinates of the representative point of the obstacle candidate area are not limited to the method of calculating and estimating the distance based on the parallax by the imaging devices installed on the left and right. A method of directly obtaining a sonic device and the like may be used. Needless to say, when a radar device, an ultrasonic device, or the like is provided, it is sufficient to install one far-infrared imaging device 1 and one near-infrared imaging device 2.

  The LSI 33 extracts the feature amount of the obstacle candidate area based on the read image data, and determines whether or not an object such as a pedestrian exists based on the extracted feature amount. When the LSI 33 determines that an object such as a pedestrian is present, the LSI 33 is a basis for the position coordinates, the feature amount, and the calculation of the representative point of the object area recognized as the object such as the pedestrian. The position coordinates of the vehicle at the time of capturing the image data are stored in the RAM 331.

  The LSI 33 reads out image data stored in the image memory 32 captured at the same vehicle position in different time zones in units of frames, and specifies an obstacle candidate region again. For example, when the image data that is the basis of the information on the recorded object area is captured at night, the image data captured in the daytime is read out. The LSI 33 estimates the position coordinates of the identified obstacle candidate area based on the parallax of the read image data, and queries the RAM 331 to determine whether the position coordinates of the object area recorded are different. to decide.

  When the LSI 33 determines that the position coordinates of the two areas are different, it determines that the object detection result is appropriate, that is, the information regarding the object area recorded in the RAM 331 is information regarding the pedestrian. If it is determined that the position coordinates are substantially the same, the LSI 33 determines that the recognition that the target object such as a pedestrian has been detected is an error, and the information on the target area recorded in the RAM 331 includes information such as the artificial structure. Judged as information on a stationary object.

  Hereinafter, detailed processing of the LSI 33 of the detection device 3 of the object detection system according to the second embodiment of the present invention will be described. FIG. 10 and FIG. 11 are flowcharts showing the procedure of the object detection information recording process of the LSI 33 of the detection apparatus 3 of the object detection system according to the second embodiment of the present invention.

  The LSI 33 acquires the position coordinates (latitude information and longitude information) of the vehicle from the vehicle position detection device 9 (step S1001), detected values by various sensors provided outside, on / off information of the headlights of the vehicle, and image data. Condition determination information such as an average value of the luminance values is acquired (step S1002), and it is determined whether a predetermined condition is satisfied based on the acquired condition determination information (step S1003). When the LSI 33 determines that the predetermined condition is satisfied (step S1003: YES), the LSI 33 acquires image data captured by the near-infrared imaging devices 2 and 2 (step S1004). If it is determined that the condition is not satisfied (step S1003: NO), the LSI 33 acquires image data captured by the far-infrared imaging devices 1 and 1 (step S1005).

  The criterion for determining whether or not the predetermined condition is satisfied is not particularly limited. For example, an illuminance sensor (not shown) for detecting the illuminance around the vehicle is installed in the vehicle, and it is determined that the condition is satisfied when the output value of the illuminance sensor is larger than a predetermined value. That is, when the output illuminance is greater than a predetermined value, it is determined that it is daytime, and image data captured by the near-infrared imaging devices 2 and 2 is acquired. When the output illuminance is less than or equal to a predetermined value, it is determined that it is nighttime, and image data captured by the far-infrared imaging devices 1 and 1 is acquired.

  Further, it may be determined that the condition is satisfied when the vehicle headlight switch is in an OFF state. In this case, when the vehicle headlight switch is on, it is determined that the vehicle is at night, and image data captured by the far-infrared imaging devices 1 and 1 is acquired. When the vehicle headlight switch is off, it is determined that it is daytime, and image data captured by the near-infrared imaging devices 2 and 2 is acquired.

  Furthermore, the average value of the luminance values of the image data captured by the near-infrared imaging devices 2 and 2 may be calculated, and it may be determined that the condition is satisfied when the calculated average value is greater than a predetermined value. . In this case, when the calculated average value is larger than the predetermined value, it is determined that it is daytime, and the image data captured by the near infrared imaging devices 2 and 2 is acquired. When the calculated average value is less than or equal to the predetermined value, it is determined that the night is night, and image data captured by the far-infrared imaging devices 1 and 1 is acquired.

  By switching the imaging device that acquires the image data in this manner, when the brightness around the vehicle is a certain level or more, for example, in the daytime, by using the visible light imaging device or the near infrared imaging devices 2 and 2, The obstacle candidate area can be specified, and the recognition result during night driving can be verified. Further, when the brightness around the vehicle is below a certain level, for example, at night, the far-infrared imaging devices 1 and 1 cannot be used to check with the visible light imaging device or the near-infrared imaging devices 2 and 2. The presence of the object can be reliably detected.

  The LSI 33 divides the image data captured by the one far-infrared imaging device 1 acquired from the image memory 32 into small regions of a predetermined size in order to identify obstacle candidate regions (step S1006).

  The LSI 33 uses the divided small area as a reference area, and designates an area corresponding to one reference area in the image data captured by one far-infrared imaging device 1, that is, a corresponding area indicating the same object as the other far-infrared imaging apparatus 1. Search is performed from image data captured by the infrared imaging device 1. Specifically, between one reference region in the image data captured by one far-infrared imaging device 1 and a search region of the same size as the reference region in the image data captured by the other far-infrared imaging device 1. Correlation calculation is performed (step S1007), and it is determined whether the correlation value is greater than a predetermined value (step S1008).

  When the LSI 33 determines that the correlation value between the reference area and the search area is smaller than the predetermined value (step S1008: NO), the LSI 33 changes the search area, returns the process to step S1007, and repeats the above-described process. When the LSI 33 determines that the correlation value is greater than the predetermined value (step S1008: YES), the LSI 33 regards the search area as a corresponding area and calculates the parallax between the reference area and the search area (step S1009).

  The LSI 33 determines whether or not the parallax has been calculated for all reference regions in the image captured by one far-infrared imaging device 1 (step S1010), and the LSI 33 determines that there is an uncalculated reference region. If it is (step S1010: NO), the LSI 33 changes the reference area (step S1011), returns the process to step S1007, and repeats the above-described process. When the LSI 33 determines that the parallax has been calculated for all the reference areas in the image (step S1010: YES), the LSI 33 combines adjacent small areas having substantially the same parallax and identifies it as an obstacle candidate area ( Step S1012).

  The LSI 33 extracts a feature amount used to determine whether or not an object such as a pedestrian exists in the specified obstacle candidate area, and the obstacle candidate area is determined to be a pedestrian or the like in accordance with a determination rule. It is determined whether or not the target area indicates the target object. As an example, a correlation value by pattern matching between pixel data of an obstacle candidate region and pixel data of a template indicating a standard human temperature distribution is calculated as a feature amount (step S1013), and a predetermined value set as a determination rule It is determined whether or not the correlation value calculated from the value is large (step S1014). When the LSI 33 determines that the correlation value is equal to or less than the predetermined value (step S1014: NO), the LSI 33 determines that the area is not an object area indicating the presence of an object such as a pedestrian, and the process returns to step S1013. The same process is repeated for the next obstacle candidate area.

  When the LSI 33 determines that the correlation value is greater than the predetermined value (step S1014: YES), the LSI 33 determines that the obstacle candidate area is an object area indicating an object such as a pedestrian and the like, and calculates The position coordinates of the representative point of the object area, the extracted feature amount, and the position coordinates of the vehicle when the image data specifying the object area are acquired are recorded in the RAM 331 (step S1015).

  12 and 13 are flowcharts showing a procedure of erroneous recognition avoidance processing of the LSI 33 of the detection device 3 of the object detection system according to Embodiment 2 of the present invention. The LSI 33 acquires the vehicle position coordinates from the vehicle position detection device 9 (step S1201), and determines whether or not the vehicle position coordinates substantially match the vehicle position coordinates recorded in the RAM 331 (step S1202). If it is determined that the LSI 33 is different (step S1202: NO), the LSI 33 returns to step S1201 and acquires the position coordinates of the vehicle. When it is determined that the LSIs 33 substantially match (step S1202: YES), the LSI 33 detects the values detected by various sensors provided outside, the on / off information of the vehicle headlights, and the average value of the luminance values of the image data. Or the like (step S1203), and based on the acquired condition determination information, it is determined whether a predetermined condition is satisfied (step S1204).

  When the LSI 33 determines that the predetermined condition is satisfied (step S1204: YES), the LSI 33 acquires the image data captured by the near-infrared imaging devices 2 and 2 (step S1205). If it is determined that the condition is not satisfied (step S1204: NO), the LSI 33 acquires image data captured by the far-infrared imaging devices 1 and 1 (step S1206).

  When the vehicle travels at the same position at different times, there is almost no obstacle candidate region having the same feature amount at the position where the presence of an object such as a pedestrian is detected last time. This is because pedestrians move. However, when detecting artificial structures such as roadside utility poles, streetlight poles, vending machines, etc. that show a temperature distribution similar to that of pedestrians, etc., even if the times taken are different, they are substantially the same. An object having substantially the same feature amount at the position of is detected.

  Therefore, when the vehicle travels at the same position at different times, and there is an obstacle candidate area having the same feature amount at the position where the presence of the object such as the previous pedestrian is detected, the object such as the previous pedestrian Judging that the presence of an object itself is an error and leaving a record of the position and feature amount of the previous object area, it will be erroneously recognized as an object such as a pedestrian after the next time. You can avoid that. Conversely, when the vehicle travels at the same position at different times and there is no obstacle candidate area having the same feature amount at the position detected as the object of the previous pedestrian or the like, the object of the previous pedestrian or the like It is determined that the presence of the object is detected correctly, and the position and feature amount of the object area recorded in the RAM 331 are deleted.

  In order to implement the above processing, the LSI 33 divides the image data acquired from the image memory 32 into small areas of a predetermined size (step S1207).

  The LSI 33 uses a divided small area as a reference area, a reference area in image data captured by one far-infrared imaging device 1, and a search area in image data captured by the other far-infrared imaging apparatus 1. A normal cross-correlation value is calculated (step S1208), and it is determined whether or not there is a search region having a correlation value greater than a predetermined value (step S1209).

  When the LSI 33 determines that there is no search area larger than the predetermined value (step S1209: NO), the LSI 33 changes the reference area, returns the process to step S1208, and repeats the above-described process. When the LSI 33 determines that there is a search area larger than the predetermined value (step S1209: YES), the LSI 33 has a parallax between the reference area and the search area, that is, one of the far infrared imaging devices 1, for example, a far disposition installed on the right. A correlation value is predetermined in an image captured by a reference region in an image captured by the infrared imaging device 1 and the other far infrared imaging device 1 corresponding to the reference region, for example, a far infrared imaging device 1 installed on the left. The parallax with the reference area larger than the value is calculated (step S1210).

  The LSI 33 determines whether or not the parallax has been calculated for all the small areas in the image (step S1211). If the LSI 33 determines that there is a small area that has not been calculated (step S1211: NO), the LSI 33 The small area is changed (step S1212), the process returns to step S1208, and the above-described process is repeated. When the LSI 33 determines that the parallax has been calculated for all the small areas in the image (step S1211: YES), the LSI 33 combines adjacent small areas having substantially the same parallax and identifies it as an obstacle candidate area ( Step S1213).

  Next, the LSI 33 estimates the position coordinates of the center of gravity (representative point) of the obstacle candidate area by estimating the distance to the obstacle candidate area based on the parallax images of the left and right imaging devices (step S1214). .

  Then, the LSI 33 determines whether or not the estimated position coordinates of the obstacle candidate area are different from the position coordinates of the object area recorded in the RAM 331 (step S1215). When it is determined that the LSI 33 is different (step S1215: YES), the LSI 33 determines that the process for detecting the presence of the object such as the pedestrian is the correct process, and the object area recorded in the RAM 331 is determined. The position coordinates, the feature amount of the image data corresponding to the object area, and the position coordinates of the vehicle when the image data specifying the object area are acquired are deleted (step S1216). Whether or not the position coordinates are different is determined based on whether or not the estimated position coordinates of the obstacle candidate area are included within a predetermined range from the position coordinates of the object area recorded in the RAM 331.

  If it is determined that the LSIs 33 substantially match (step S1215: NO), the LSI 33 determines that the process of detecting the presence of an object such as a pedestrian is incorrect last time and that the detected object is an artificial structure. The record of the object area in the RAM 331 is left (step S1217). From the next time, the position coordinates of the object area recorded in the RAM 331, the feature amount of the image data corresponding to the object area, and the position coordinates of the vehicle at the time of obtaining the image data specifying the object area are collated. By doing so, it becomes possible to avoid erroneously recognizing an object that is not an object such as a pedestrian but an artificial structure as an object such as a pedestrian. Therefore, the LSI 33 does not output a signal indicating that the presence of an object such as a pedestrian has been detected to the display device 4 or the alarm device 5, and the driver is bothered by unnecessary warning display, alarm sound, and the like. You can drive safely without any problems.

As described above, according to the second embodiment, when an object area that has been recognized as an object such as a pedestrian is not an object such as a pedestrian at a separate timing, The position coordinates of the vehicle that has determined that the recognition itself as an object such as a pedestrian is an error, and acquired the position coordinates, the feature amount of the object area, and the image data specifying the object area Can be prevented from being erroneously recognized as an object such as a pedestrian in the subsequent object detection process. In addition, it is not necessary for the driver to register information consciously, and it is possible to continue driving safely.
In the second embodiment, the case where the far-infrared imaging device and the near-infrared imaging device are provided has been described. However, two imaging devices having different wavelength bands that can be imaged may be provided. For example, any two of a visible light imaging device, a far infrared imaging device, and a near infrared imaging device may be provided. However, it is more preferable to provide the far-infrared imaging device and the near-infrared imaging device because there is a low possibility of erroneously recognizing a stationary object other than a pedestrian or the like that emits infrared rays.

  Other configurations, operations, and effects of the object detection system according to the second embodiment are the same as the configurations, operations, and effects of the object detection system according to the first embodiment. The detailed description is omitted.

It is a schematic diagram which shows the structure of the target object detection system which concerns on Embodiment 1 of this invention. It is a block diagram which shows the structure of the far-infrared imaging device of the target object detection system which concerns on Embodiment 1 of this invention. It is a block diagram which shows the structure of the detection apparatus of the target object detection system which concerns on Embodiment 1 of this invention. It is a flowchart which shows the procedure of the recording process of the target object detection information of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 1 of this invention. It is a flowchart which shows the procedure of the recording process of the target object detection information of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 1 of this invention. It is a flowchart which shows the procedure of the misrecognition avoidance process of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 1 of this invention. It is a flowchart which shows the procedure of the misrecognition avoidance process of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 1 of this invention. It is a schematic diagram which shows the structure of the target object detection system which concerns on Embodiment 2 of this invention. It is a block diagram which shows the structure of the detection apparatus of the target object detection system which concerns on Embodiment 2 of this invention. It is a flowchart which shows the procedure of the recording process of the target object detection information of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 2 of this invention. It is a flowchart which shows the procedure of the recording process of the target object detection information of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 2 of this invention. It is a flowchart which shows the procedure of the misrecognition avoidance process of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 2 of this invention. It is a flowchart which shows the procedure of the misrecognition avoidance process of LSI of the detection apparatus of the target object detection system which concerns on Embodiment 2 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Far-infrared imaging device 2 Near-infrared imaging device 3 Detection apparatus 4 Display apparatus 5 Alarm apparatus 6 Car-mounted LAN cable 7 Video cable 8 Cable 9 Vehicle position detection apparatus 31a Video input part 31b Video output part 31c Communication interface part 32 Image memory 33 LSI
331 RAM

Claims (6)

In an object detection system comprising: an imaging device that captures the periphery of a vehicle; and a detection device that acquires an image captured by the imaging device and detects an object included in the image;
A vehicle position detection device for detecting vehicle position information;
The detection device includes:
Obstacle candidate area specifying means for specifying an obstacle candidate area based on one image data acquired from the imaging device;
Means for determining whether or not the object region corresponds to the object based on the feature amount of the one image data corresponding to the identified obstacle candidate region;
Means for recording the position coordinates of the object area and the position information of the vehicle at the time when one image data specifying the object area is acquired, when the means determines that it is the object area;
Means for acquiring other image data at a position corresponding to the recorded vehicle position information at a different time from the time when the one image data was acquired;
Means for respecifying an obstacle candidate area based on the other image data;
Means for determining whether or not the position coordinates of the re-specified obstacle candidate area and the position coordinates of the recorded object area are substantially the same;
Means for deleting the position coordinates of the object area recorded and the position information of the vehicle at the time of acquiring one image data specifying the object area when it is determined that the means does not substantially match And
An object detection system characterized in that erroneous recognition of an object is prevented based on the recorded position coordinates of the object area and vehicle position information. In an object detection system comprising: a first imaging device that images the periphery of a vehicle; and a detection device that acquires an image captured by the first imaging device and detects an object included in the image.
A vehicle position detecting device for detecting position information of the vehicle;
A second imaging device having a wavelength band that can be imaged different from that of the first imaging device,
The detection device includes:
Obstacle candidate area specifying means for specifying an obstacle candidate area based on one image data acquired from the first imaging device;
Means for determining whether or not the object region corresponds to the object based on the feature amount of the one image data corresponding to the identified obstacle candidate region;
Means for recording the position coordinates of the object area and the position information of the vehicle at the time when one image data specifying the object area is acquired, when the means determines that it is the object area;
Means for selecting the second imaging device if the predetermined condition is satisfied, and selecting the first imaging device if the predetermined condition is not satisfied;
At a time point different from the time point when the first image pickup device acquires one image data at a position corresponding to the vehicle position information recorded by the selected first image pickup device or second image pickup device. Means for acquiring image data of
Means for respecifying an obstacle candidate area based on the other image data;
Means for determining whether or not the position coordinates of the re-specified obstacle candidate area and the position coordinates of the recorded object area are substantially the same;
Means for deleting the position coordinates of the object area recorded and the position information of the vehicle at the time of acquiring one image data specifying the object area when it is determined that the means does not substantially match And
An object detection system characterized in that erroneous recognition of an object is prevented based on the recorded position coordinates of the object area and vehicle position information. The first imaging device is a far infrared imaging device, and the second imaging device is a visible light imaging device or a near infrared imaging device,
It has an illuminance sensor that detects ambient illuminance,
The criterion for determining whether or not the predetermined condition is satisfied is whether or not the detection value of the illuminance sensor exceeds a predetermined value,
When it is determined that the detection value of the illuminance sensor exceeds a predetermined value, the second imaging device is selected, and when it is determined that the detection value is not more than the predetermined value, the first imaging device is selected. The object detection system according to claim 2, wherein there is an object detection system. The first imaging device is a far infrared imaging device, and the second imaging device is a visible light imaging device or a near infrared imaging device,
It has a switch to turn on and off the vehicle headlight,
The criterion for determining whether or not the predetermined condition is satisfied is whether or not the switch is in an off state,
The second imaging device is selected when it is determined that the switch is in an off state, and the first imaging device is selected when it is determined that the switch is in an on state. The object detection system according to claim 2. The first imaging device is a far infrared imaging device, and the second imaging device is a visible light imaging device or a near infrared imaging device,
The criterion for determining whether or not the predetermined condition is satisfied is whether or not an average value of luminance values of images captured by the second imaging device exceeds a predetermined value.
When it is determined that the average value of luminance values exceeds a predetermined value, the second imaging device is selected, and when it is determined that the average value is equal to or less than the predetermined value, the first imaging device is selected. The object detection system according to claim 2, wherein the object detection system is configured as described above. In an object detection method for acquiring an image captured by an imaging device that captures the periphery of a vehicle and detecting an object included in the image,
Identifying an obstacle candidate area based on one image data acquired from the imaging device,
It is determined whether or not the object area corresponds to the object based on the feature amount of the one image data corresponding to the identified obstacle candidate area,
When it is determined that the object area, the position coordinates of the object area and the position information of the vehicle at the time when one image data specifying the object area is acquired are recorded,
At a time point different from the time point when the one image data was acquired, at a position corresponding to the recorded vehicle position information, acquire other image data,
Re-specify the obstacle candidate area based on the other image data,
Determine whether the position coordinates of the re-specified obstacle candidate area and the position coordinates of the recorded object area substantially match,
If it is determined that they do not substantially match, the position coordinates of the recorded object area and the position information of the vehicle at the time of acquiring one image data specifying the object area are deleted, Target detection method.

Images