JP4768499B2 - In-vehicle peripheral other vehicle detection device - Google Patents
In-vehicle peripheral other vehicle detection device Download PDFInfo
- Publication number
- JP4768499B2 JP4768499B2 JP2006113108A JP2006113108A JP4768499B2 JP 4768499 B2 JP4768499 B2 JP 4768499B2 JP 2006113108 A JP2006113108 A JP 2006113108A JP 2006113108 A JP2006113108 A JP 2006113108A JP 4768499 B2 JP4768499 B2 JP 4768499B2
- Authority
- JP
- Japan
- Prior art keywords
- vehicle
- motorcycle
- camera
- region
- wheeled
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Image Processing (AREA)
- Closed-Circuit Television Systems (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Description
本発明は、自動車に搭載される、自動車周辺の他車を検出する車載周辺他車検出装置に関するものである。 The present invention relates to a vehicle-mounted peripheral other vehicle detection device that detects other vehicles around the vehicle and is mounted on the vehicle.
自動車周辺の他車を検出する技術としては、カメラで撮影した自車周辺の画像に対して画像認識処理を施して、自車周辺に存在する他車を検出し、検出した他車の位置などの情報を提示する技術が知られている（たとえば、特許文献１）。
運転者の周辺状況認識の支援のために、自車近くに存在する自動二輪車を検出し運転者に対して提示しようとする場合、前記カメラで撮影した自車周辺の画像に対する画像認識処理によって他車を識別する技術によれば次のような問題が生じる。
すなわち、自車近くの自動二輪車は、自車近くの他の自動四輪車の陰となってカメラで撮影できなくなってしまうことが多い。そして、このために、自車近くの自動二輪車の検出漏れが多く発生する。
一方で、自車に近づいてくる自動二輪車を、自車から遠い位置にある時点から検出するようにし、過去に検出した自動二輪車位置やその推移から、現在、当該自動二輪車が自車の近くに存在するかどうかを推定することにより、自車近くの自動二輪車を検出するようにすることも考えられる。しかし、このようにすると、カメラで撮影した画像の広い領域について画像認識処理を行わなければならない。また、遠距離にあって画像中に写り込むサイズの小さい自動二輪車を認識するために、高精度な画像認識処理を行わなければならなくなる。すなわち、自車から遠い位置にある自動二輪車を検出するようにすると、画像認識処理の処理量が増大し、効率的な他車検出の妨げとなる。
In order to assist the driver in recognizing the surrounding situation of the driver, when a motorcycle near the own vehicle is detected and presented to the driver, the image recognition processing is performed on the image around the own vehicle taken by the camera. According to the technology for identifying a car, the following problems occur.
That is, a motorcycle near the own vehicle is often behind the other automobiles near the own vehicle and cannot be photographed by the camera. For this reason, many detection omissions occur in motorcycles near the host vehicle.
On the other hand, a motorcycle approaching the own vehicle is detected from a point far from the own vehicle, and the motorcycle is currently located near the own vehicle from the position and transition of the motorcycle detected in the past. It is also conceivable to detect a motorcycle near the own vehicle by estimating whether it exists. However, if this is done, image recognition processing must be performed for a wide area of the image taken by the camera. In addition, in order to recognize a small motorcycle that is at a long distance and appears in the image, it is necessary to perform highly accurate image recognition processing. That is, if a motorcycle far from the own vehicle is detected, the amount of image recognition processing increases, which hinders efficient detection of other vehicles.
そこで、本発明は、カメラで撮影した画像に基づいて、より効率的に自車周辺の他車を適正に検出することを課題とする。 Therefore, an object of the present invention is to appropriately detect other vehicles around the own vehicle more efficiently based on an image photographed by a camera.
前記課題達成のために、本発明は、自動車に搭載され、前記自動車を自車として自車周辺の他車を検出する車載周辺他車検出装置に、自車周辺を撮影するカメラと、前記カメラが撮影した画像に対して画像認識処理を施して自車周辺の他車を検出する画像認識処理部とを備えると共に、前記画像認識処理部を、前記カメラが撮影した画像の、自車に対して最も近い領域である第１の領域が写り込む部分である第１の部分に対して、自動四輪車と自動二輪車の位置を検出する画像認識処理を施す近距離他車認識手段と、前記カメラが撮影した画像の、前記第１の領域の次に自車に対して近い領域である第２の領域が写り込む部分である第２の部分に対して、自動四輪車の位置を検出する画像認識処理を施す遠距離四輪車認識手段と、前記遠距離四輪車認識手段が、前記自動四輪車の位置を検出した場合に、前記カメラが撮影した画像の、当該検出された自動四輪車の周辺の領域である第３の領域が写り込む部分である第３の部分に対して、自動二輪輪車の位置を検出することのできる画像認識処理を施す遠距離二輪車認識手段とより構成したものである。 In order to achieve the above object, the present invention provides a vehicle-mounted peripheral other vehicle detection device that is mounted on a vehicle and detects other vehicles around the vehicle using the vehicle as the own vehicle, An image recognition processing unit that performs image recognition processing on an image captured by the vehicle and detects other vehicles around the vehicle, and the image recognition processing unit is configured to apply the image captured by the camera to the vehicle. Short-range other vehicle recognition means for performing image recognition processing for detecting the positions of the four-wheeled motor vehicle and the two-wheeled motor vehicle with respect to the first part that is a part in which the first region which is the closest region is reflected, The position of the four-wheeled vehicle is detected with respect to the second part of the image captured by the camera, which is the part where the second area, which is the area closest to the own vehicle next to the first area, is reflected. A long-distance four-wheeled vehicle recognition means for performing image recognition processing, and the long-distance When the wheel recognizing means detects the position of the automobile, the third area of the image taken by the camera is the area around the detected automobile. A third part is constituted by a long-distance motorcycle recognition means for performing image recognition processing capable of detecting the position of the motorcycle.
ここで、自車から遠い位置にある自動二輪車のうち、自車に接近したときに他の自動四輪車の陰になる可能性の高い自動二輪車は、少なくとも他の自動四輪車の周辺に位置する自動二輪車である。したがって、本車載周辺他車検出装置のように、自車から遠い位置に自動四輪車が存在する場合にのみ、当該自動四輪車の周辺を対象として自車から遠い位置にある自動二輪車を検出するための画像認識処理を行うことにより、当該画像認識処理の処理量を抑えつつ、自車に接近したときに他の自動四輪車の陰になる可能性のある自動二輪車を、当該自動二輪車が自車から遠い位置にある時点から検出することができる。そして、このようにして過去に検出した自動二輪車位置やその推移から、現在、当該自動二輪車が自車の近くに存在するかどうかを推定することにより、自車近くの自動二輪車を当該自動二輪車が他の自動四輪車の死角に位置する場合も適正に検出するようができるようになる。 Here, among motorcycles that are far from the vehicle, motorcycles that are likely to be in the shadow of other motorcycles when approaching the vehicle are at least around other motorcycles. The motorcycle is located. Therefore, only when an automobile is present at a position far from the own vehicle as in this on-vehicle surrounding other vehicle detection device, a motorcycle at a position far from the own vehicle is targeted around the automobile. By performing image recognition processing for detection, a motorcycle that may be shaded by other automobiles when approaching the own vehicle is suppressed while the processing amount of the image recognition processing is suppressed. It can be detected from the time when the two-wheeled vehicle is far from the vehicle. Then, by estimating whether the motorcycle is present near the own vehicle from the position of the motorcycle detected in the past and its transition, the motorcycle near the own vehicle is Even when the vehicle is located in the blind spot of another automobile, it can be properly detected.
ここで、より具体的には、車載周辺他車検出装置において、前記第３の領域としては、たとえば、遠距離四輪車認識手段が位置を検出した前記自動四輪車の、前記カメラから見て左右に位置する領域を用いることができる。または、前記第３の領域としては、遠距離四輪車認識手段が位置を検出した前記自動四輪車が自車に接近したときに、当該自動四輪車によって生じる前記カメラの死角となる領域と同じ相対位置関係を、当該自動四輪車に対して有する領域を用いるようにしてもよい。 More specifically, in the on-vehicle surrounding other vehicle detection device, the third region is, for example, a view of the automatic four-wheeled vehicle detected by the long-distance four-wheeled vehicle recognition means from the camera. Thus, regions located on the left and right can be used. Alternatively, as the third region, a region that becomes a blind spot of the camera generated by the four-wheeled vehicle when the four-wheeled vehicle whose position is detected by the long-distance four-wheeled vehicle recognition means approaches the own vehicle. A region having the same relative positional relationship with respect to the automobile may be used.
これらのようにすることにより、自車から遠い位置にある自動二輪車を検出するための画像認識処理の対象とする画像の部分を、現在遠い位置にあって自車に接近したときに他の自動四輪車の陰になる可能性がより高い自動二輪車が存在する部分に限定することができ、この結果、当該画像認識処理の処理量を増大をより抑制することができるようになる。 By doing in this way, the part of the image subject to image recognition processing for detecting a motorcycle far from the own vehicle is moved to another automatic when it is in the current far position and approaches the own vehicle. This can be limited to a portion where there is a motorcycle more likely to be shaded by a four-wheeled vehicle, and as a result, the processing amount of the image recognition processing can be further suppressed.
ここで、以上のような車載周辺他車検出装置は、前記カメラは前記自動車後方を撮影するものとして良い。また、この場合には、車載周辺他車検出装置に、前記遠距離四輪車認識手段が前記自動四輪車の位置を検出した場合に、当該検出された自動四輪車の位置が、自車と同じ走行レーン上にあるかどうかを算定する走行レーン算定手段を設け、前記遠距離二輪車認識手段において、前記走行レーン算定手段が、前記遠距離四輪車認識手段が検出した自動四輪車の位置が自車と同じ走行レーン上にあることを算定した場合にのみ、前記第３の部分に対して、自動二輪輪車の位置を検出することのできる画像認識処理を施すようにしてもよい。 Here, in the on-vehicle surrounding other vehicle detection device as described above, the camera may take an image of the rear of the vehicle. In this case, when the long-range four-wheel vehicle recognition means detects the position of the four-wheeled vehicle, the detected position of the four-wheeled vehicle is A travel lane calculating means for calculating whether or not the vehicle is on the same travel lane as the vehicle, and in the long-distance motorcycle recognition means, the travel lane calculation means is an automatic four-wheeled vehicle detected by the long-distance four-wheel vehicle recognition means. Only when it is calculated that the position of the vehicle is on the same travel lane as the vehicle, an image recognition process capable of detecting the position of the motorcycle is performed on the third portion. Good.
ここで、現在自車後方の遠い位置にあって自車に接近したときに他の自動四輪車の陰になる可能性の高い自動二輪車は、自車と同じ走行レーンを走行している他の自動四輪車の周辺に位置する自動二輪車である。したがって、このように、自車と同じ走行レーンを走行している他の自動四輪車が存在する場合にのみ、自車から遠い位置にある自動二輪車を検出するための画像認識処理を行うことにより、当該画像認識処理の処理量を増大をより抑えることができる。 Here, a motorcycle that is currently in a far position behind the host vehicle and that is likely to be behind the other four-wheeler when approaching the host vehicle is traveling in the same lane as the host vehicle. This is a motorcycle located around the automobile. Therefore, the image recognition process for detecting a motorcycle at a position far from the own vehicle is performed only when there is another four-wheeled vehicle traveling in the same traveling lane as the own vehicle. As a result, the increase in the processing amount of the image recognition processing can be further suppressed.
また、以上のような車載周辺他車検出装置には、前記遠距離二輪車認識手段が過去に検出した自動二輪車の位置に基づいて、自車近傍の自動四輪車のために前記カメラの死角となっている範囲内に存在する自動二輪車の位置を推定する二輪車位置推定手段を設けることが好ましい。このようにすることにより、他の自動四輪車の死角に存在する自車周辺の自動二輪車を検出して、検出した自動二輪車についての情報を提示したりすることができるようになる。すなわち、この場合には、たとえば、他の自動四輪車の死角に存在する自動二輪車についての情報を提示する手段として、前記二輪車位置推定手段が推定した自動二輪車の位置が、自車の状態と所定の関係を有するものであった場合に、当該自動二輪車の存在を提示する状況提示手段などを備えることができる。 In addition, in the above-mentioned on-vehicle surrounding other vehicle detection device, the blind spot of the camera for the four-wheeled vehicle in the vicinity of the own vehicle based on the position of the motorcycle detected by the long-distance motorcycle recognition unit in the past. It is preferable to provide a motorcycle position estimating means for estimating the position of the motorcycle existing within the range. By doing so, it is possible to detect a motorcycle around the own vehicle existing in the blind spot of another motorcycle and present information about the detected motorcycle. That is, in this case, for example, as a means for presenting information about a motorcycle existing in the blind spot of another motorcycle, the position of the motorcycle estimated by the motorcycle position estimating means is the state of the own vehicle. When the vehicle has a predetermined relationship, it can be provided with a situation presenting means for presenting the presence of the motorcycle.
以上のように、本発明によれば、カメラで撮影した画像に基づいて、より効率的に自車周辺の他車を適正に検出することができる。 As described above, according to the present invention, it is possible to more efficiently detect other vehicles around the host vehicle based on the image taken by the camera.
以下、本発明の実施形態について説明する。
図１ａに、本実施形態に係る周辺他車提示装置の構成を示す。
本周辺状況他車装置は自動車に搭載される装置であり、図示するように、後方カメラ１、画像認識処理部２、他車追尾部３、周辺状況提示部４、車両状態センサ５、表示装置７、音声出力装置７とを備えている。ここで、車両状態センサ５は、自動車の方向指示器操作状態などの車両の各種状態を計測するセンサの集合である。
Hereinafter, embodiments of the present invention will be described.
FIG. 1 a shows a configuration of a nearby other vehicle presentation device according to the present embodiment.
This peripheral state other vehicle device is a device mounted on a car, and as shown in the figure, a rear camera 1, an image recognition processing unit 2, another vehicle tracking unit 3, a peripheral state presenting unit 4, a vehicle state sensor 5, and a display device. 7 and an audio output device 7. Here, the vehicle state sensor 5 is a set of sensors that measure various states of the vehicle such as the direction indicator operation state of the automobile.
また、図１ｂに示すように、後方カメラ１は自動車の前部に設置され、自車の後方を撮影する。
また、画像認識処理部２は、以下に示す画像認識処理を行う。
図２ａに、この画像認識処理の手順を示す。
図示するように、この処理では、まず、後方カメラ１が撮影した画像の近距離認識領域を対象とする、近距離四輪車及び二輪車認識処理を起動する（ステップ２０２）。ここで、近距離認識領域は、後方カメラ１が撮影した画像中の、自車に対して所定距離以内の被写体が写り込む領域であり、たとえば、図３ａに示すように、後方カメラ１が撮影した画像３００の下方の領域３０１として予め設定する。
Further, as shown in FIG. 1b, the rear camera 1 is installed in the front part of the automobile and photographs the rear of the own vehicle.
The image recognition processing unit 2 performs the following image recognition process.
FIG. 2a shows the procedure of this image recognition process.
As shown in the figure, in this process, first, a short-distance four-wheeled vehicle and two-wheeled vehicle recognition process for a short-distance recognition region of an image taken by the rear camera 1 is started (step 202). Here, the short distance recognition area is an area in which an object within a predetermined distance appears in the image taken by the rear camera 1, and for example, the rear camera 1 takes a picture as shown in FIG. 3a. An area 301 below the image 300 is set in advance.
ここで、起動された近距離四輪車及び二輪車認識処理は、所定の近距離四輪車認識用の画像認識アルゴリズムを用いた、後方カメラ１が撮影した画像の近距離認識領域内に存在する四輪車の画像パターンの検出と、検出した画像パターンの画像中の位置より推定される当該四輪車の自車に対する相対位置の算定とを繰り返し行う。また、起動された近距離四輪車及び二輪車認識処理は、所定の近距離二輪車認識用の画像認識アルゴリズムを用いた、後方カメラ１が撮影した画像の近距離認識領域内に存在する二輪車の画像パターンの検出と、検出した画像パターンの画像中の位置より推定される当該二輪車の自車に対する相対位置の算定とを繰り返し行う。また、近距離四輪車及び二輪車認識処理では、算定した四輪車の相対位置を、四輪車の現在位置データとして他車追尾部３に出力し、算定した二輪車の相対位置を、二輪車の現在位置データとして他車追尾部３に出力する。 Here, the activated short-range four-wheeled vehicle and two-wheeled vehicle recognition processing exists in the short-range recognition region of the image captured by the rear camera 1 using a predetermined image recognition algorithm for short-range four-wheeled vehicle recognition. The detection of the image pattern of the four-wheel vehicle and the calculation of the relative position of the four-wheel vehicle with respect to the own vehicle estimated from the position of the detected image pattern in the image are repeated. The activated short-range four-wheeled vehicle and two-wheeled vehicle recognition processing uses a predetermined short-range two-wheeled vehicle recognition image recognition algorithm, and the image of the two-wheeled vehicle existing in the short-range recognition region of the image captured by the rear camera 1. The detection of the pattern and the calculation of the relative position of the two-wheeled vehicle estimated from the position of the detected image pattern in the image are repeated. In the short-range four-wheeled and two-wheeled vehicle recognition processing, the calculated relative position of the four-wheeled vehicle is output to the other vehicle tracking unit 3 as the current position data of the four-wheeled vehicle, and the calculated relative position of the two-wheeled vehicle is The current position data is output to the other vehicle tracking unit 3.
図２ａに戻り、次に、後方カメラ１が撮影した画像の遠距離四輪車認識領域を対象とする、遠距離四輪車認識処理を起動する（ステップ２０４）。ここで、遠距離四輪車認識領域は、近距離認識領域３０１に写り込む被写体よりも自車に対してある程度遠い被写体が写り込む、後方カメラ１が撮影した画像中の領域であり、たとえば、図３ａに示すように、後方カメラ１が撮影した画像３００の、近距離認識領域３０１の上部の領域３０２として予め設定する。 Returning to FIG. 2a, next, a long-distance four-wheeled vehicle recognition process for the long-distance four-wheeled vehicle recognition region of the image taken by the rear camera 1 is started (step 204). Here, the long-distance four-wheel vehicle recognition area is an area in an image taken by the rear camera 1 in which a subject that is somewhat distant from the subject vehicle is reflected in the short-distance recognition area 301. As shown in FIG. 3 a, it is set in advance as an area 302 above the short-distance recognition area 301 of the image 300 taken by the rear camera 1.
ここで、起動された遠距離四輪車認識処理は、所定の遠距離四輪車認識用の画像認識アルゴリズムを用いた、後方カメラ１が撮影した画像の遠距離四輪車認識領域内の四輪車の画像パターンの検出と、検出した画像パターンの画像中の位置より推定される当該四輪車の自車に対する相対位置の算定とを繰り返し行う。なお、遠距離四輪車認識処理では、距離認識領域内の二輪車の認識は行わない。また、遠距離四輪車認識処理では、現在位置データの他車追尾部３への出力を行わない。 Here, the activated long-distance four-wheel vehicle recognition process uses four images in the long-distance four-wheel vehicle recognition area of the image captured by the rear camera 1 using a predetermined image recognition algorithm for long-distance four-wheel vehicle recognition. The detection of the image pattern of the wheel vehicle and the calculation of the relative position of the four-wheel vehicle with respect to the own vehicle estimated from the position of the detected image pattern in the image are repeated. In the long-distance four-wheel vehicle recognition process, the two-wheeled vehicle in the distance recognition area is not recognized. In the long-distance four-wheel vehicle recognition process, the current position data is not output to the other vehicle tracking unit 3.
図２に戻り、以上のようにして、近距離四輪車及び二輪車認識処理と遠距離四輪車認識処理を起動したならば、遠距離四輪車認識処理による四輪車の認識の発生、すなわち、遠距離四輪車認識処理による遠距離四輪車認識領域内の四輪車の画像パターンの検出の発生を監視する（ステップ２０６）。 Returning to FIG. 2, if the short-range four-wheeled vehicle and two-wheeled vehicle recognition processing and the long-range four-wheeled vehicle recognition processing are started as described above, the recognition of the four-wheeled vehicle by the long-range four-wheeled vehicle recognition processing, That is, the detection of the detection of the image pattern of the four-wheel vehicle in the long-distance four-wheel vehicle recognition area by the long-distance four-wheel vehicle recognition process is monitored (step 206).
そして、遠距離四輪車認識処理による四輪車の認識が発生したならば、その認識された四輪車が、自車の後方において自車と同じ走行レーン（車線）を走行している他車であるかどうかを識別し（ステップ２０８）、自車と同じ走行レーンを走行中でなければ（ステップ２１０）、ステップ２０６に戻って、遠距離四輪車認識処理による次の四輪車の認識の発生を待つ。ここで、遠距離四輪車認識処理によって認識された四輪車が自車の後方において自車と同じ走行レーン（車線）を走行している他車であるかどうかは、たとえば、遠距離四輪車認識処理が算定した当該四輪車の自車に対する相対位置や、後方カメラ１で撮影した画像に対して白線認識処理を施して認識した自車後方の走行レーンの配置に応じて識別するようにする。 If the recognition of a four-wheeled vehicle by the long-distance four-wheeled vehicle recognition processing occurs, the recognized four-wheeled vehicle is traveling in the same lane (lane) as the own vehicle behind the own vehicle. If the vehicle is not traveling in the same traveling lane as that of the own vehicle (step 210), the process returns to step 206 to determine the next four-wheeled vehicle by the long-distance four-wheeled vehicle recognition processing. Wait for recognition to occur. Here, whether the four-wheeled vehicle recognized by the long-distance four-wheeled vehicle recognition process is another vehicle traveling behind the own vehicle in the same traveling lane (lane) as the own vehicle is, for example, Identification is made according to the relative position of the four-wheeled vehicle calculated by the wheel-recognition processing and the arrangement of the driving lane behind the vehicle recognized by performing white line recognition processing on the image captured by the rear camera 1. Like that.
そして、遠距離四輪車認識処理によって認識された四輪車が自車と同じ走行レーンを走行中であれば（ステップ２１０）。後方カメラ１で撮影した画像中に、遠距離二輪車認識領域を設定し（ステップ２１２）、設定した遠距離二輪車認識領域を対象とする、遠距離二輪車認識処理を起動し（ステップ２１４）、ステップ２０６に戻って、遠距離四輪車認識処理による次の四輪車の認識の発生を待つ。 If the four-wheeled vehicle recognized by the long-distance four-wheeled vehicle recognition process is traveling in the same traveling lane as the own vehicle (step 210). A long-distance motorcycle recognition area is set in the image captured by the rear camera 1 (step 212), and a long-distance motorcycle recognition process for the set long-distance motorcycle recognition area is started (step 214). Returning to, it waits for the occurrence of recognition of the next four-wheel vehicle by the long-distance four-wheel vehicle recognition processing.
ここで、ステップ２１２において、遠距離二輪車認識領域は、次のように設定する。すなわち、たとえば、図３ｂのように、後方カメラ１が撮影した画像３００中の遠距離四輪車認識領域３０２内に、四輪車の画像パターン３０２１が検出されたならば、図３ｃに示すように、その画像パターン３０２１の左右に、それぞれ遠距離二輪車認識領域３０３を設定する。 Here, in step 212, the long-distance motorcycle recognition area is set as follows. That is, for example, as shown in FIG. 3B, if a four-wheeled vehicle image pattern 3021 is detected in a long-distance four-wheeled vehicle recognition area 302 in the image 300 taken by the rear camera 1, as shown in FIG. 3C. In addition, a long-distance motorcycle recognition area 303 is set on each of the left and right sides of the image pattern 3021.
遠距離二輪車領域３０３のサイズは、図３ｄに示すように、画像パターン３０２１の下辺上の左右方向について中央の位置Ｘを基準に定める。
すなわち、二つの遠距離二輪車領域３０３の水平方向のサイズは、自車後方に水平な地面のみが存在するものとして次のように求める。すなわち、後方カメラ１が撮影した画像３００の位置Ｘに写り込む地面上の地点から、自車左方向に所定距離Ｄ移動した地点が写り込むことになる後方カメラ１が撮影した画像３００中の位置Ａが、左側の遠距離二輪車領域３０３の左下角の点となり、後方カメラ１が撮影した画像３００の位置Ｘに写り込む地面上の地点から、自車右方向に所定距離Ｄ移動した地点が写り込むことになる後方カメラ１が撮影した画像３００中の位置Ｂが、右側の遠距離二輪車領域３０３の右下角の点となるように二つの遠距離二輪車領域３０３のサイズは定める。ここで、このように二つの遠距離二輪車領域３０３の水平方向のサイズを求めるために用いる、所定距離Ｄは、１走行レーンの幅に所定のオフセットを加えた長さの半分の長さとする。なお、このように二つの遠距離二輪車領域３０３の水平方向のサイズを定める結果、左側の遠距離二輪車領域３０３の左下角の点Ａと、右側の遠距離二輪車領域３０３の右下角の点Ｂの距離Ｗは、画像パターン３０２１が認識された他車の位置する地点を中心とする、１走行レーンの幅に所定のオフセットを加えた左右方向の長さに対応するものとなる。
The size of the long-distance motorcycle region 303 is determined based on the center position X in the horizontal direction on the lower side of the image pattern 3021 as shown in FIG.
That is, the horizontal size of the two long-distance motorcycle regions 303 is obtained as follows assuming that only the horizontal ground exists behind the host vehicle. That is, the position in the image 300 taken by the rear camera 1 that the point moved by the predetermined distance D in the left direction of the vehicle is reflected from the point on the ground reflected in the position X of the image 300 taken by the rear camera 1. A is a point at the lower left corner of the far-distance motorcycle region 303 on the left side, and a point moved by a predetermined distance D in the right direction from the point on the ground reflected in the position X of the image 300 taken by the rear camera 1 is shown. The size of the two long-distance motorcycle regions 303 is determined so that the position B in the image 300 captured by the rear camera 1 to be included is the lower right corner point of the long-distance motorcycle region 303 on the right side. Here, the predetermined distance D used for obtaining the horizontal size of the two long-distance two-wheeled vehicle regions 303 in this way is a half of the length obtained by adding a predetermined offset to the width of one traveling lane. In addition, as a result of determining the horizontal size of the two long-distance motorcycle regions 303 in this way, the lower left corner point A of the left long-distance motorcycle region 303 and the lower right corner point B of the right long-distance motorcycle region 303 The distance W corresponds to the length in the left-right direction obtained by adding a predetermined offset to the width of one traveling lane centered on the point where the other vehicle where the image pattern 3021 is recognized.
次に、二つの遠距離二輪車領域の垂直方向のサイズＨは、自車後方に水平な地面のみが存在するものとして次のように求める。すなわち、後方カメラ１が撮影した画像３００の位置Ｘに写り込む地面上の地点に配置した、自動二輪車の高さに所定のオフセットを加えた高さを有する物体の上端が写り込むことになる後方カメラ１が撮影した画像３００中の位置Ｃと、位置Ｘと間の距離を遠距離二輪車領域の垂直方向のサイズＨとして求める。 Next, the vertical size H of the two long-distance motorcycle regions is obtained as follows assuming that only the horizontal ground exists behind the vehicle. That is, the upper end of an object having a height obtained by adding a predetermined offset to the height of the motorcycle, which is arranged at a position on the ground that is reflected in the position X of the image 300 taken by the rear camera 1, is reflected. The distance between the position C in the image 300 captured by the camera 1 and the position X is obtained as the vertical size H of the long-distance motorcycle region.
以上、画像認識処理によって説明した。
次に、この画像認識処理のステップ２１４で起動する遠距離二輪車認識処理について説明する。
図２ｂに示すように、この処理では、所定の遠距離二輪車認識用の画像認識アルゴリズムを用いて、後方カメラ１が撮影した画像の遠距離二輪車認識領域内に存在する二輪車の画像パターンを検出する（ステップ２５２）。そして、検出に失敗した場合には（ステップ２５４）、そのまま処理を終了する。一方、検出に成功した場合には（ステップ２５４）、検出した画像パターンの画像中の位置より推定される当該二輪車の自車に対する相対位置を算定し、算定した二輪車の相対位置を、二輪車の現在位置データとして他車追尾部３に出力する（ステップ２５６）。そして、遠距離二輪車認識処理を終了する。
The image recognition process has been described above.
Next, the long-distance motorcycle recognition process activated in step 214 of the image recognition process will be described.
As shown in FIG. 2b, in this process, the image pattern of the two-wheeled vehicle existing in the long-distance motorcycle recognition area of the image captured by the rear camera 1 is detected by using a predetermined image recognition algorithm for long-distance motorcycle. (Step 252). If the detection fails (step 254), the process is terminated as it is. On the other hand, if the detection is successful (step 254), the relative position of the two-wheeled vehicle estimated from the position in the image of the detected image pattern is calculated, and the calculated relative position of the two-wheeled vehicle is calculated as the current position of the two-wheeled vehicle. The position data is output to the other vehicle tracking unit 3 (step 256). Then, the long-distance motorcycle recognition process ends.
ここで、以上のような画像認識処理の処理例を示す。
いま、図４ａに示すような配置で、自車４０１の後方に他の四輪車４０２と自動二輪車４０３が存在する場合を考える。
この場合、図４ｂの自車４０１に近い範囲４１１が、後方カメラ１によって近距離認識領域内の画像として撮影され、近距離認識領域を対象として他車認識を行う近距離四輪車及び二輪車認識処理によって、当該範囲４１１に存在する四輪車と二輪車の認識が行われる。一方、範囲４１１の次に自車に近い範囲４１２は、後方カメラ１によって遠距離四輪車認識領域内の画像として撮影され、遠距離四輪車認識領域を対象として四輪車の認識を行う遠距離四輪車認識処理によって、当該範囲４１２に存在する四輪車認識が行われる。そして、この結果、範囲４１２内に存在する四輪車４０２が、遠距離四輪車認識処理によって認識される。
Here, a processing example of the image recognition processing as described above will be shown.
Consider a case where another four-wheeled vehicle 402 and a motorcycle 403 are present behind the host vehicle 401 in the arrangement as shown in FIG.
In this case, the range 411 close to the own vehicle 401 in FIG. 4B is captured as an image in the short-range recognition area by the rear camera 1, and the short-range four-wheeled vehicle and the two-wheeled vehicle recognition that performs other vehicle recognition on the short-range recognition area as a target. By the processing, the recognition of the four-wheeled vehicle and the two-wheeled vehicle existing in the range 411 is performed. On the other hand, the range 412 closest to the host vehicle next to the range 411 is taken by the rear camera 1 as an image in the long-distance four-wheel vehicle recognition area, and the four-wheeled vehicle is recognized for the long-distance four-wheel vehicle recognition area. The four-wheeled vehicle existing in the range 412 is recognized by the long-distance four-wheeled vehicle recognition process. As a result, the four-wheeled vehicle 402 existing within the range 412 is recognized by the long-distance four-wheeled vehicle recognition process.
また、四輪車４０２が遠距離四輪車認識処理によって認識されると、図４ｃに示すように、四輪車４０２の周辺の四輪車４０２に対して左右および斜め後ろ方向に位置する範囲４１３が写り込む範囲が、遠距離二輪車認識領域として設定される。なお、前述のように遠距離二輪車認識領域のサイズが設定されるため、この範囲４１３の、他車４０２の自車４０１側端における左右方向の幅ＷＲは、走行レーンの幅に所定のオフセットを加えたものとなる。 In addition, when the four-wheeled vehicle 402 is recognized by the long-distance four-wheeled vehicle recognition processing, as shown in FIG. 4C, a range located in the left and right and obliquely rearward directions with respect to the four-wheeled vehicle 402 around the four-wheeled vehicle 402. A range in which 413 is reflected is set as a long-distance motorcycle recognition area. In addition, since the size of the long-distance motorcycle recognition area is set as described above, the width WR in the left-right direction at the vehicle 401 side end of the other vehicle 402 in this range 413 has a predetermined offset to the width of the travel lane. It will be added.
ここで、このように遠距離二輪車認識領域が設定されると、遠距離二輪車認識領域を対象として二輪車の認識を行う遠距離二輪車認識処理が行われ、この結果、範囲４１３内に存在する二輪車４０３が、遠距離四輪車認識処理によって認識されることになる。
以上、画像認識処理部２の動作について説明した。
ところで、以上の画像認識処理で用いる遠距離二輪車認識領域は、次のように設定するようにしてもよい。
すなわち、図５ａに示すように、（図４ｂと同様に）遠距離四輪車認識領域の画像として撮影される範囲４１２内に存在する四輪車４０２が、遠距離四輪車認識処理によって認識されたならば、この四輪車４０２が、図５ｂに示すようにそのまま自車に接近した場合、当該接近した四輪車４０２１によって後方カメラ１の死角となる相対範囲５０１を、当該接近した四輪車４０２１に対して求める。そして、図５ｃに示すように、当該相対範囲５０１と接近した四輪車４０２１との位置関係と同じ位置関係を、他車４０２に対して有する範囲５０２を求める。
Here, when the long-distance motorcycle recognition area is set in this way, a long-distance motorcycle recognition process for recognizing the two-wheeled vehicle is performed for the long-distance motorcycle recognition area. As a result, the two-wheeled vehicle 403 existing in the range 413 is obtained. Are recognized by the long-distance four-wheel vehicle recognition process.
The operation of the image recognition processing unit 2 has been described above.
By the way, the long-distance motorcycle recognition area used in the above image recognition processing may be set as follows.
That is, as shown in FIG. 5a, the four-wheeled vehicle 402 existing in the range 412 photographed as an image of the long-distance four-wheeled vehicle recognition area is recognized by the long-distance four-wheeled vehicle recognition processing (as in FIG. 4b). If the four-wheeled vehicle 402 approaches the vehicle as it is, as shown in FIG. 5b, the relative range 501 that is the blind spot of the rear camera 1 by the approaching four-wheeled vehicle 4021 It asks for wheel set 4021. Then, as shown in FIG. 5 c, a range 502 having the same positional relationship with respect to the other vehicle 402 as the positional relationship between the relative range 501 and the approaching four-wheeled vehicle 4021 is obtained.
そして、図５ｄに示すように、範囲５０２を含む、対応する画像上の領域が矩形となる最小の範囲５０３が写り込む範囲を遠距離二輪車認識領域として設定する。
以上のように、本実施形態に係る画像認識処理部２では、自車から遠い位置にある二輪車を検出するために行う遠距離二輪車認識処理を、現在自車から遠い位置にあって自車に接近したときに、後方カメラ１に対して自車周辺の二輪車を隠す死角を作る可能性の高い四輪車が存在する場合にのみ、当該四輪車が自車に接近したときに当該四輪車の陰となる二輪車が現在存在する可能性の高い領域のみを対象として行う。したがって、遠距離二輪車認識処理の処理量を抑えつつ、現在自車から遠い位置にあって自車に接近したときに他の四輪車の陰になる可能性のある自動二輪車を適正に検出することができる。
Then, as shown in FIG. 5d, a range including a range 502 in which a minimum range 503 in which a corresponding region on the image is rectangular is reflected is set as a long-distance motorcycle recognition region.
As described above, in the image recognition processing unit 2 according to the present embodiment, the long-distance motorcycle recognition processing performed to detect the two-wheeled vehicle at a position far from the own vehicle is performed on the own vehicle at a position far from the own vehicle. When there is a four-wheeled vehicle that has a high possibility of creating a blind spot that hides the two-wheeled vehicle around the rear vehicle 1 when approaching, the four-wheeled vehicle approaches the four-wheeled vehicle when approaching the own vehicle. This is done only in areas where there is a high possibility that a motorcycle behind the car will be present. Therefore, while suppressing the processing amount of the long-distance motorcycle recognition processing, a motorcycle that is currently at a position far from the own vehicle and that may be shaded by other four-wheel vehicles when approaching the own vehicle is properly detected. be able to.
さて、図１に戻り、他車追尾部３は、画像認識処理部２から四輪車及び二輪車の現在位置データに基づいて、自車後方の各他車（四輪車及び二輪車）の追尾を行い、追尾中の各他車毎に、追尾情報を生成し管理する他車追尾処理を繰り返し行う。ここで、他車の追尾とは、各時点で画像認識処理部２から出力された現在位置データを、個々の他車毎に、当該他車の時系列の現在位置データとして系列化する処理である。また、ある他車の追尾情報とは、このように系列化された当該他車の時系列の現在位置データを指す。すなわち、他車追尾部３によって、同じ他車（四輪車または二輪車）に対して各時点で画像認識処理部２で算定された各現在位置データは、一つの他車の追尾情報として管理される。なお、一般的には、画像認識処理部２から出力される現在位置データのうちに、ある追尾中の他車の過去の現在位置データから推定される当該他車の現在位置近傍の現在位置を示す現在位置データが存在する場合に、当該現在位置データを、当該追尾中の他車の現在位置データであるとすることにより、当該追尾中の他車の追尾を行う。また、画像認識処理部２から出力される現在位置データのうちに、追尾中の他車の現在位置データとすることのできなかった現在位置データが存在する場合には、新たに追尾する他車を生成し、当該存在した現在位置データを、生成した追尾する他車の現在位置データとする。ただし、以上の他車の追尾の処理は、画像認識処理部２が出力する四輪車の現在位置データと、画像認識処理部２が出力する二輪車現在位置データについて各々独立的に行う。 Now, referring back to FIG. 1, the other vehicle tracking unit 3 tracks the other vehicles (four-wheeled vehicles and two-wheeled vehicles) behind the own vehicle based on the current position data of the four-wheeled vehicles and the two-wheeled vehicles from the image recognition processing unit 2. The other vehicle tracking process for generating and managing the tracking information is repeated for each other vehicle being tracked. Here, the tracking of the other vehicle is a process of serializing the current position data output from the image recognition processing unit 2 at each time point as time-series current position data of the other vehicle for each other vehicle. is there. Further, the tracking information of a certain other vehicle refers to the time-series current position data of the other vehicle thus arranged. That is, the current position data calculated by the image recognition processing unit 2 at each time for the same other vehicle (four-wheeled vehicle or two-wheeled vehicle) by the other vehicle tracking unit 3 is managed as tracking information of one other vehicle. The In general, the current position data output from the image recognition processing unit 2 includes the current position in the vicinity of the current position of the other vehicle estimated from the past current position data of the other vehicle being tracked. When the current position data shown is present, the current position data is the current position data of the other vehicle being tracked, thereby tracking the other vehicle being tracked. If the current position data output from the image recognition processing unit 2 includes the current position data that could not be the current position data of the other vehicle being tracked, the other vehicle to be newly tracked And the existing current position data is used as the current position data of the other vehicle to be tracked. However, the tracking processing of other vehicles described above is performed independently for the current position data of the four-wheeled vehicle output by the image recognition processing unit 2 and the current position data of the two-wheeled vehicle output by the image recognition processing unit 2.
ここで、他車追尾部３は、最近に行われた近距離四輪車及び二輪車認識処理による近距離認識領域の画像認識によっても、最近に行われた遠距離二輪車認識処理によっても、対応する現在位置データが出力されなかった追尾中の他車については、その追尾を終了し、当該他車についての追尾情報を消去する。ただし、対応する現在位置データが出力されなかった他車が二輪車である場合には、次に示す死角範囲内二輪車位置補完処理を行って、当該二輪車の現在の現在位置データの補完を試み、補完が成功した場合には、補完した現在位置データを用いて、当該二輪車の追尾を継続する。 Here, the other vehicle tracking unit 3 responds by the image recognition of the short-distance recognition area by the short-range four-wheeled vehicle and the two-wheeled vehicle recognition processing performed recently or by the long-range two-wheeled vehicle recognition processing performed recently. For other vehicles that are being tracked for which no current position data has been output, the tracking is terminated and the tracking information for the other vehicles is deleted. However, if the other vehicle for which the corresponding current position data is not output is a two-wheeled vehicle, the following blind spot two-wheeler position supplement processing is performed to try to complement the current current position data of the motorcycle. If is successful, the tracking of the motorcycle is continued using the supplemented current position data.
図６に、この死角範囲内二輪車位置補完処理の手順を示す。
図示するように、この処理では、対応する現在位置データが出力されなかった追尾中の二輪車について以下の処理を行う。
すなわち、まず、当該追尾中の二輪車について、最近の過去所定期間内に、現在位置データを画像認識処理部２より取得できた回数が所定のしきい値Ｔｈより大きいかどうかを当該追尾中の二輪車の追尾情報を参照して調べ（ステップ６０２）、回数がしきい値Ｔｈ以下であれば、補完失敗として、そのまま処理を終了する。
FIG. 6 shows the procedure of the blind spot range two-wheeled vehicle position complementing process.
As shown in the figure, in this processing, the following processing is performed for a tracking two-wheeled vehicle for which corresponding current position data has not been output.
That is, first, regarding the two-wheeled vehicle being tracked, it is determined whether the number of times the current position data can be acquired from the image recognition processing unit 2 within a recent past predetermined period is greater than a predetermined threshold Th. The tracking information is checked (step 602), and if the number of times is equal to or less than the threshold value Th, the processing is terminated as it is as a complement failure.
一方、最近の過去所定期間内に、現在位置データを画像認識処理部２より取得できた回数が所定のしきい値Ｔｈより大きい場合には、当該追尾中の二輪車の追尾情報が示すこれまでの現在位置データの推移に基づいて、当該二輪車の現在の相対位置を推定する（ステップ６０４）。そして、推定した相対位置が、他の四輪車や二輪車によって後方カメラ１に対して死角となっている範囲内にあるかどうかを、各追尾中他車の追尾情報を参照して調べ（ステップ６０６）、死角となっている範囲内になければ、補完失敗としてそのまま処理を終了する。 On the other hand, when the number of times the current position data can be acquired from the image recognition processing unit 2 within the recent past predetermined period is larger than the predetermined threshold value Th, the tracking information of the two-wheeled vehicle being tracked so far is indicated. Based on the transition of the current position data, the current relative position of the motorcycle is estimated (step 604). Then, it is checked whether or not the estimated relative position is within a range that is a blind spot with respect to the rear camera 1 by another four-wheeled vehicle or two-wheeled vehicle with reference to the tracking information of the other vehicles during each tracking (step 606) If it is not within the range of the blind spot, the process is terminated as it is as a complement failure.
一方、推定した現在位置が死角となっている範囲内にあれば、推定した相対位置を表す現在位置データを、当該二輪車の現在の現在位置データとして、当該二輪車の追尾情報中に補完し（ステップ６０８）、処理を終了する。なお、現在位置データを追尾情報中に補完する場合には、当該補完した現在位置データが画像認識処理部２より取得したものではなく、補完したものであることを示す情報も、追尾情報中に含めるようにする。このようにすることにより、当該二輪車について後に行われる死角範囲内二輪車位置補完処理のステップ６０２において、追尾情報より、最近の過去所定期間内に現在位置データを画像認識処理部２より取得できた回数を適正に求めることができるようになる。 On the other hand, if the estimated current position is within the blind spot, the current position data representing the estimated relative position is complemented as the current current position data of the two-wheeled vehicle in the tracking information of the two-wheeled vehicle (step 608), the process is terminated. When the current position data is complemented in the tracking information, the information indicating that the supplemented current position data is not acquired from the image recognition processing unit 2 but is supplemented is also included in the tracking information. To include. In this way, the number of times the current position data can be acquired from the image recognition processing unit 2 within the most recent past predetermined period from the tracking information in step 602 of the blind spot in-range two-wheeler position complementing process performed later for the two-wheeled vehicle. Can be obtained appropriately.
以上、他車追尾部３の動作について説明した。
このように、本実施形態に係る他車追尾部３によれば、自車周辺の二輪車が、他の四輪車の死角に存在する場合であっても、当該二輪車の現在の相対位置を、過去に前述した遠距離二輪車認識処理で求めた二輪車の現在位置データを用いて推定、検出することができる。
次に、図１において、周辺状況提示部４は、他車追尾部３が追尾中の各他車の現在の相対位置に基づいて、周辺状況を提示する処理を行う。
この周辺状況の処理では、たとえば、自車を表す図形と共に、他車追尾部３が追尾中の各他車を表すマークを、自車を表す図形に対して、当該他車の追尾情報の最新の相対位置が示す当該他車と自車との位置関係に従って配置した画像を表示装置７に表示することにより、各他車の現在の相対位置を提示する。また、追尾中のいずれかの他車の追尾情報が示す現在の相対位置が、所定のレベルを超えて他車が自車に接近していることを表している場合に、その旨を警告する音声を音声出力装置７から出力する。また、車両状態センサ５が検出している方向指示器操作状態より自車の進路変更が予測されたときに、追尾中のいずれかの他車の追尾情報が示す現在の相対位置が、自車後方の自車の進路変更予定方向側に他車が存在していることを表している場合に、その旨を警告する音声を音声出力装置７から出力する処理なども行う。
The operation of the other vehicle tracking unit 3 has been described above.
Thus, according to the other vehicle tracking unit 3 according to the present embodiment, even if the two-wheeled vehicle around the own vehicle exists in the blind spot of the other four-wheeled vehicle, the current relative position of the two-wheeled vehicle is It can be estimated and detected using the current position data of the two-wheeled vehicle obtained by the above-described long-distance two-wheeled vehicle recognition process.
Next, in FIG. 1, the surrounding situation presentation unit 4 performs a process of presenting the surrounding situation based on the current relative position of each other vehicle being tracked by the other vehicle tracking unit 3.
In the processing of the surrounding situation, for example, together with a graphic representing the own vehicle, a mark representing each other vehicle being tracked by the other vehicle tracking unit 3 is used to update the tracking information of the other vehicle with respect to the graphic representing the own vehicle. The current relative position of each other vehicle is presented by displaying an image arranged according to the positional relationship between the other vehicle and the host vehicle indicated by the relative position on the display device 7. In addition, if the current relative position indicated by the tracking information of any other vehicle being tracked exceeds a predetermined level and indicates that the other vehicle is approaching the vehicle, a warning to that effect is given. Audio is output from the audio output device 7. Further, when the course change of the own vehicle is predicted from the direction indicator operation state detected by the vehicle state sensor 5, the current relative position indicated by the tracking information of any other vehicle being tracked is When it is shown that there is another vehicle on the side of the planned direction change of the rear vehicle, the voice output device 7 outputs a warning to that effect.
なお、ここで、このように周辺状況の提示処理で取り扱う追尾中の他車には、前記死角範囲内二輪車位置補完処理で現在位置が補完された自動二輪車も含まれる。したがって、このような周辺状況提示部４の動作によれば、自車周辺の他の四輪車の死角に存在する二輪車についても、その情報を提示することができる。 It should be noted that the other vehicles being tracked that are handled in the processing for presenting the surrounding situation as described above include a motorcycle in which the current position is complemented by the motorcycle position supplement processing in the blind spot range. Therefore, according to the operation of the surrounding state presenting unit 4 as described above, it is possible to present information about a two-wheeled vehicle existing in the blind spot of another four-wheeled vehicle around the own vehicle.
以上、周辺状況提示部４の処理について説明した。
以上、本発明の実施形態について説明した。
ところで、以上の実施形態では、次の動作を行うようにしてもよい。
すなわち、上述した画像認識処理において、遠距離四輪車認識処理で四輪車Ｘが認識され、かつ、当該認識に応答して設定された遠距離二輪車認識領域に対する遠距離二輪車認識処理で二輪車Ｙが認識されたならば、遠距離四輪車認識処理で求めた四輪車Ｘの現在位置データを、遠距離二輪車認識処理で認識した二輪車Ｙの遠距離四輪車認識処理で認識した四輪車Ｘに対する相対位置を伴わせた形態で、他車追尾部３に出力する。
The processing of the surrounding situation presentation unit 4 has been described above.
The embodiment of the present invention has been described above.
Incidentally, in the above embodiment, the following operation may be performed.
That is, in the image recognition process described above, the four-wheeled vehicle X is recognized in the long-distance four-wheeled vehicle recognition process, and the two-wheeled vehicle Y in the long-distance two-wheeled vehicle recognition process for the long-distance two-wheeled vehicle recognition region set in response to the recognition. Is recognized, the current position data of the four-wheeled vehicle X obtained in the long-distance four-wheeled vehicle recognition processing is recognized by the long-distance four-wheeled vehicle recognition processing of the two-wheeled vehicle Y recognized in the long-distance two-wheeled vehicle recognition processing. The data is output to the other vehicle tracking unit 3 in a form accompanied by a relative position with respect to the vehicle X.
また、以降、他車追尾部３は、遠距離四輪車認識処理から出力される現在位置データと近距離四輪車及び二輪車認識処理から出力される現在位置データに基づいて四輪車Ｘの追尾を行う。そして、四輪車Ｘが自車に接近したならば、遠距離四輪車認識処理から最後に出力された四輪車Ｘの現在位置データに伴っていた二輪車Ｙの相対位置と同じ位置関係を四輪車Ｘに対して持つ位置が、現在の四輪車Ｘによる後方カメラ１の死角範囲に含まれるかどうかを判定し、死角範囲に含まれている場合には、当該位置に二輪車Ｙが存在するものと推定する。 In addition, the other vehicle tracking unit 3 thereafter determines the position of the four-wheeled vehicle X based on the current position data output from the long-distance four-wheeled vehicle recognition process and the current position data output from the short-range four-wheeled and two-wheeled vehicle recognition processing. Perform tracking. And if the four-wheeled vehicle X approaches the own vehicle, the same positional relationship as the relative position of the two-wheeled vehicle Y accompanying the current position data of the four-wheeled vehicle X outputted last from the long-distance four-wheeled vehicle recognition processing is obtained. It is determined whether or not the position of the four-wheeled vehicle X is included in the blind spot range of the rear camera 1 by the current four-wheeled vehicle X. If the position is included in the blind spot range, the two-wheeled vehicle Y is in that position. Presumed to exist.
そして、周辺状況提示部４において、推定した二輪車Ｙの位置に基づく周辺状況の提示処理を行う。すなわち、たとえば、車両状態センサ５が検出している方向指示器操作状態より自車の進路変更が予測されたときに、自車後方の、自車の進路変更予定方向側に、推定した二輪車Ｙの位置が存在している場合には、その旨を警告する音声を音声出力装置７から出力する処理などを行う。 Then, the surrounding situation presentation unit 4 performs a surrounding situation presentation process based on the estimated position of the two-wheeled vehicle Y. That is, for example, when the course change of the own vehicle is predicted from the direction indicator operation state detected by the vehicle state sensor 5, the estimated two-wheeled vehicle Y is placed on the direction of the planned course change of the own vehicle behind the own vehicle. If there is a position, the voice output device 7 performs a process of outputting a voice warning to that effect.
このように本実施形態によれば、自車から遠い位置にある二輪車を検出するための画像認識処理の処理量を抑えつつ、現在自車から遠い位置にあって自車に接近したときに他の四輪車の陰になる可能性のある二輪車を検出することができる。また、このようにして過去に行った二輪車の検出内容に基づいて、自車周辺の他の四輪車の死角範囲内に存在する二輪車を検出し、検出した二輪車についての情報を提示することができる。 As described above, according to the present embodiment, the amount of image recognition processing for detecting a two-wheeled vehicle at a position far from the own vehicle is suppressed, while the other is detected when approaching the own vehicle at a position far from the own vehicle. It is possible to detect two-wheeled vehicles that may be behind the four-wheeled vehicle. In addition, based on the detection contents of the two-wheeled vehicle performed in the past, the two-wheeled vehicle existing in the dead angle range of the other four-wheeled vehicle around the own vehicle can be detected and information on the detected two-wheeled vehicle can be presented. it can.
１…後方カメラ、２…画像認識処理部、３…他車追尾部、４…周辺状況提示部、５…車両状態センサ、７…音声出力装置、７…表示装置。 DESCRIPTION OF SYMBOLS 1 ... Back camera, 2 ... Image recognition process part, 3 ... Other vehicle tracking part, 4 ... Peripheral condition presentation part, 5 ... Vehicle state sensor, 7 ... Audio | voice output apparatus, 7 ... Display apparatus.
Claims (8)
自車周辺を撮影するカメラと、
前記カメラが撮影した画像に対して画像認識処理を施して自車周辺の他車を検出する画像認識処理部とを有し、
前記画像認識処理部は、
前記カメラが撮影した画像の、自車に対して最も近い領域である第１の領域が写り込む部分である第１の部分に対して、自動四輪車と自動二輪車の位置を検出する画像認識処理を施す近距離他車認識手段と、
前記カメラが撮影した画像の、前記第１の領域の次に自車に対して近い領域である第２の領域が写り込む部分である第２の部分に対して、自動四輪車の位置を検出する画像認識処理を施す遠距離四輪車認識手段と、
前記遠距離四輪車認識手段が、前記自動四輪車の位置を検出した場合に、前記カメラが撮影した画像の、当該検出された自動四輪車の周辺の領域である第３の領域が写り込む部分である第３の部分に対して、自動二輪車の位置を検出することのできる画像認識処理を施す遠距離二輪車認識手段とを有し、
前記第３の領域は、遠距離四輪車認識手段が位置を検出した前記自動四輪車の、前記カメラから見て左右に位置する領域であることを特徴とする車載周辺他車検出装置。 A vehicle-mounted peripheral other vehicle detection device that is mounted on an automobile and detects other vehicles around the own vehicle as the vehicle.
A camera that captures the area around the vehicle,
An image recognition processing unit that performs image recognition processing on an image captured by the camera and detects other vehicles around the own vehicle;
The image recognition processing unit
Image recognition for detecting the positions of an automobile and a motorcycle with respect to a first portion of the image taken by the camera, which is a portion where a first region that is the closest region to the own vehicle is reflected. Short-range other vehicle recognition means for processing,
The position of the four-wheeled vehicle is determined with respect to the second part of the image captured by the camera, which is the part where the second area, which is the area closest to the own vehicle next to the first area, is reflected. A long-distance automobile recognition means for performing image recognition processing to detect;
When the long-distance four-wheel vehicle recognition means detects the position of the four-wheeled vehicle, a third region that is a region around the detected four-wheeled vehicle of the image taken by the camera is for the third portion which is visible on captured moiety, it possesses a long-range motorcycle recognition means for carrying out image recognition processing which can detect the position of the motorcycle,
The on- vehicle peripheral other vehicle detection device, wherein the third region is a region located on the left and right of the four-wheeled vehicle detected by the long-distance four-wheel vehicle recognition unit as viewed from the camera .
自車周辺を撮影するカメラと、
前記カメラが撮影した画像に対して画像認識処理を施して自車周辺の他車を検出する画像認識処理部とを有し、
前記画像認識処理部は、
前記カメラが撮影した画像の、自車に対して最も近い領域である第１の領域が写り込む部分である第１の部分に対して、自動四輪車と自動二輪車の位置を検出する画像認識処理を施す近距離他車認識手段と、
前記カメラが撮影した画像の、前記第１の領域の次に自車に対して近い領域である第２の領域が写り込む部分である第２の部分に対して、自動四輪車の位置を検出する画像認識処理を施す遠距離四輪車認識手段と、
前記遠距離四輪車認識手段が、前記自動四輪車の位置を検出した場合に、前記カメラが撮影した画像の、当該検出された自動四輪車の周辺の領域である第３の領域が写り込む部分である第３の部分に対して、自動二輪車の位置を検出することのできる画像認識処理を施す遠距離二輪車認識手段とを有し、
前記第３の領域は、遠距離四輪車認識手段が位置を検出した前記自動四輪車が自車に接近したときに、当該自動四輪車によって生じる前記カメラの死角となる領域と同じ相対位置関係を、当該自動四輪車に対して有する領域であることを特徴とする車載周辺他車検出装置。 A vehicle-mounted peripheral other vehicle detection device that is mounted on an automobile and detects other vehicles around the own vehicle as the vehicle.
A camera that captures the area around the vehicle,
An image recognition processing unit that performs image recognition processing on an image captured by the camera and detects other vehicles around the own vehicle;
The image recognition processing unit
Image recognition for detecting the positions of an automobile and a motorcycle with respect to a first portion of the image taken by the camera, which is a portion where a first region that is the closest region to the own vehicle is reflected. Short-range other vehicle recognition means for processing,
The position of the four-wheeled vehicle is determined with respect to the second part of the image captured by the camera, which is the part where the second area, which is the area closest to the own vehicle next to the first area, is reflected. A long-distance automobile recognition means for performing image recognition processing to detect;
When the long-distance four-wheel vehicle recognition means detects the position of the four-wheeled vehicle, a third region that is a region around the detected four-wheeled vehicle of the image taken by the camera is Long-distance motorcycle recognition means for performing image recognition processing capable of detecting the position of the motorcycle with respect to the third portion, which is a portion to be reflected,
The third region is the same relative to a region that becomes a blind spot of the camera generated by the four-wheeled vehicle when the four-wheeled vehicle whose position is detected by the long-distance four-wheeled vehicle recognition means approaches the own vehicle. An on- vehicle peripheral other vehicle detection device characterized by being a region having a positional relationship with respect to the automobile.
前記カメラは前記自動車後方を撮影することを特徴とする車載周辺他車検出装置。The on-vehicle surrounding other vehicle detection device, wherein the camera captures the rear of the vehicle.
前記遠距離四輪車認識手段が前記自動四輪車の位置を検出した場合に、当該検出された自動四輪車の位置が、自車と同じ走行レーン上にあるかどうかを算定する走行レーン算定手段を有し、A travel lane for calculating whether the position of the detected automobile is on the same travel lane as the own vehicle when the long-distance automobile recognition means detects the position of the automobile. Have a calculation means,
前記遠距離二輪車認識手段は、前記走行レーン算定手段が、前記遠距離四輪車認識手段が検出した自動四輪車の位置が自車と同じ走行レーン上にあることを算定した場合にのみ、前記第３の部分に対して、前記自動二輪車の位置を検出することのできる画像認識処理を施すことを特徴とする車載周辺他車検出装置。The long-distance motorcycle recognition means is only when the travel lane calculation means calculates that the position of the four-wheeled vehicle detected by the long-distance four-wheel vehicle recognition means is on the same travel lane as the own vehicle, An on-vehicle peripheral other vehicle detection device that performs image recognition processing capable of detecting the position of the motorcycle on the third portion.
前記遠距離二輪車認識手段が過去に前記第３の部分に対して検出した自動二輪車の位置の推移に基づいて、自車近傍の自動四輪車のために前記カメラの死角となっている範囲内に現在位置する自動二輪車の存在及び当該自動二輪車の位置を推定する二輪車位置推定手段を、さらに有することを特徴とする車載周辺他車検出装置。Based on the transition of the position of the motorcycle detected by the long-distance motorcycle recognition unit with respect to the third portion in the past, within a range that is a blind spot of the camera for the motorcycle near the vehicle. An on-vehicle peripheral other vehicle detection device further comprising: a motorcycle position estimation means for estimating the presence of a motorcycle currently located and a position of the motorcycle.
前記二輪車位置推定手段が推定した自動二輪車の位置が、自車後方の、自車の方向指示器の操作状態から予測される進路変更方向側の位置であった場合に、当該自動二輪車の存在を提示する状況提示手段を有することを特徴とする車載周辺他車検出装置。If the position of the motorcycle estimated by the motorcycle position estimation means is a position on the side of the course change direction that is predicted from the operation state of the direction indicator of the own vehicle, the presence of the motorcycle is determined. An on-vehicle surrounding other vehicle detection device characterized by having status presenting means for presenting.
前記装置が、カメラによって自車周辺を撮影するステップと、
前記装置が、前記カメラが撮影した画像の、自車に対して最も近い領域である第１の領域が写り込む部分である第１の部分に対して、自動四輪車と自動二輪車の位置を検出する画像認識処理を施すステップと、
前記装置が、前記カメラが撮影した画像の、前記第１の領域の次に自車に対して近い領域である第２の領域が写り込む部分である第２の部分に対して、自動四輪車の位置を検出する画像認識処理を施すステップと、
前記装置が、前記第２の部分に対する画像認識処理によって前記自動四輪車が検出された場合に、前記カメラが撮影した画像の、当該検出された自動四輪車の周辺の領域である第３の領域が写り込む部分である第３の部分に対して、自動二輪車の位置を検出することのできる画像認識処理を施すステップとを有し、
前記第３の領域は、遠距離四輪車認識手段が位置を検出した前記自動四輪車の、前記カメラから見て左右に位置する領域であることを特徴とする周辺他車検出方法。 In a device mounted on an automobile, a surrounding other vehicle detection method for detecting other vehicles around the own vehicle as the own vehicle,
The step of photographing the surroundings of the vehicle by a camera;
The position of the four-wheeled motor vehicle and the two-wheeled motor vehicle with respect to the first portion in which the first region which is the closest region to the own vehicle in the image captured by the camera is reflected. Performing an image recognition process to detect;
An automatic four-wheel vehicle for a second portion in which an image captured by the camera is a portion in which a second region that is closest to the host vehicle next to the first region is reflected. Performing image recognition processing for detecting the position of the car;
When the apparatus detects the four-wheeled vehicle by an image recognition process for the second portion, a third area that is an area around the detected four-wheeled vehicle of the image captured by the camera Performing an image recognition process capable of detecting the position of the motorcycle on the third part, which is a part in which the area of
The third area is a surrounding other vehicle detection method, which is an area located on the left and right of the four-wheeled vehicle detected by the long-distance four-wheel vehicle recognition means as viewed from the camera .
前記装置が、カメラによって自車周辺を撮影するステップと、
前記装置が、前記カメラが撮影した画像の、自車に対して最も近い領域である第１の領域が写り込む部分である第１の部分に対して、自動四輪車と自動二輪車の位置を検出する画像認識処理を施すステップと、
前記装置が、前記カメラが撮影した画像の、前記第１の領域の次に自車に対して近い領域である第２の領域が写り込む部分である第２の部分に対して、自動四輪車の位置を検出する画像認識処理を施すステップと、
前記装置が、前記第２の部分に対する画像認識処理によって前記自動四輪車が検出された場合に、前記カメラが撮影した画像の、当該検出された自動四輪車の周辺の領域である第３の領域が写り込む部分である第３の部分に対して、自動二輪車の位置を検出することのできる画像認識処理を施すステップとを有し、
前記第３の領域は、遠距離四輪車認識手段が位置を検出した前記自動四輪車が自車に接近したときに、当該自動四輪車によって生じる前記カメラの死角となる領域と同じ相対位置関係を、当該自動四輪車に対して有する領域であることを特徴とする周辺他車検出方法。 In a device mounted on an automobile, a surrounding other vehicle detection method for detecting other vehicles around the own vehicle as the own vehicle,
The step of photographing the surroundings of the vehicle by a camera;
The apparatus comprising an image in which the camera is taken, the first portion is a first region bleeds through portion which is closest to the area relative to the vehicle, the position of the four-wheeled vehicle and a motorcycle Performing an image recognition process to detect;
An automatic four-wheel vehicle for a second portion in which an image captured by the camera is a portion in which a second region that is closest to the host vehicle next to the first region is reflected. Performing image recognition processing for detecting the position of the car;
When the apparatus detects the four-wheeled vehicle by an image recognition process for the second portion, a third area that is an area around the detected four-wheeled vehicle of the image captured by the camera for the third portion which is the portion bleeds through the region, possess a step of performing image recognition processing which can detect the position of the motorcycle,
The third region is the same relative to a region that becomes a blind spot of the camera generated by the four-wheeled vehicle when the four-wheeled vehicle whose position is detected by the long-distance four-wheeled vehicle recognition means approaches the own vehicle. A surrounding other vehicle detection method characterized by being a region having a positional relationship with respect to the automobile .
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2006113108A JP4768499B2 (en) | 2006-04-17 | 2006-04-17 | In-vehicle peripheral other vehicle detection device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2006113108A JP4768499B2 (en) | 2006-04-17 | 2006-04-17 | In-vehicle peripheral other vehicle detection device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JP2007286873A JP2007286873A (en) | 2007-11-01 |
| JP4768499B2 true JP4768499B2 (en) | 2011-09-07 |
Family
ID=38758587
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP2006113108A Expired - Fee Related JP4768499B2 (en) | 2006-04-17 | 2006-04-17 | In-vehicle peripheral other vehicle detection device |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP4768499B2 (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2016170663A (en) | 2015-03-13 | 2016-09-23 | 株式会社Ｊｖｃケンウッド | Vehicle monitoring device, vehicle monitoring method, and vehicle monitoring program |
| KR101930884B1 (en) * | 2017-09-29 | 2019-03-11 | 충북대학교 산학협력단 | Forward vehicle detection apparatus and operation method thereof |
| JP7052632B2 (en) * | 2018-08-10 | 2022-04-12 | トヨタ自動車株式会社 | Peripheral display device for vehicles |
| CN112703541B (en) * | 2018-09-17 | 2022-09-20 | 日产自动车株式会社 | Vehicle behavior prediction method and vehicle behavior prediction device |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0973529A (en) * | 1995-07-05 | 1997-03-18 | Suzuki Motor Corp | Two-wheeled vehicle monitor device |
| JPH11108661A (en) * | 1997-08-05 | 1999-04-23 | Honda Motor Co Ltd | Distance measuring apparatus for vehicle |
| JP2000127850A (en) * | 1998-10-22 | 2000-05-09 | Matsushita Electric Ind Co Ltd | Vehicle outside information detecting device |
| JP3900814B2 (en) * | 2000-09-29 | 2007-04-04 | 三菱自動車工業株式会社 | Driving support system |
| JP2003123196A (en) * | 2001-10-10 | 2003-04-25 | Denso Corp | Device and program for monitoring circumference of vehicle |
| JP3916958B2 (en) * | 2002-01-24 | 2007-05-23 | クラリオン株式会社 | Vehicle rear monitoring system and monitoring device |
| JP2004159268A (en) * | 2002-11-08 | 2004-06-03 | Nissan Motor Co Ltd | Vehicle periphery monitoring device |
| JP4021344B2 (en) * | 2003-03-03 | 2007-12-12 | 富士重工業株式会社 | Vehicle driving support device |
| JP4720386B2 (en) * | 2005-09-07 | 2011-07-13 | 株式会社日立製作所 | Driving assistance device |
-
2006
- 2006-04-17 JP JP2006113108A patent/JP4768499B2/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| JP2007286873A (en) | 2007-11-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8289189B2 (en) | Camera system for use in vehicle parking | |
| JP4420011B2 (en) | Object detection device | |
| US7190281B2 (en) | Vehicle environment monitoring device, vehicle environment monitoring method, control program and computer-readable recording medium | |
| JP4561863B2 (en) | Mobile body path estimation device | |
| JP6190758B2 (en) | Object recognition device and vehicle | |
| JP4615038B2 (en) | Image processing device | |
| CN109204136B (en) | Peripheral image display control device | |
| JP6073358B2 (en) | Active warning and / or navigation assistance method for avoiding collision between vehicle body and / or wheel and object | |
| JP2007228448A (en) | Imaging environment recognition apparatus | |
| JP2010019589A (en) | Inter-vehicle distance detector, drive recorder apparatus | |
| US20170028917A1 (en) | Driving assistance device and driving assistance method | |
| JP2017147629A (en) | Parking position detection system, and automatic parking system using the same | |
| JP2019087875A (en) | Periphery monitoring device | |
| JP2009075638A (en) | Vehicle type discrimination device | |
| JP5097681B2 (en) | Feature position recognition device | |
| JP4768499B2 (en) | In-vehicle peripheral other vehicle detection device | |
| KR20200059267A (en) | Techniques to assist the vehicle in driving when approaching the speed breaker | |
| JP4661602B2 (en) | Rear vehicle analysis device and collision prediction device | |
| JP5832850B2 (en) | Lane monitoring system and lane monitoring method | |
| JP2010018080A (en) | Vehicular driving support device | |
| CN112542060B (en) | Rear side alarm device for vehicle | |
| JP7279390B2 (en) | electronic controller | |
| JP2006182108A (en) | Vehicle surroundings monitoring apparatus | |
| JP2017124751A (en) | Parking support device and parking support method | |
| WO2022202780A1 (en) | Display control device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A621 | Written request for application examination |
Free format text: JAPANESE INTERMEDIATE CODE: A621 Effective date: 20090220 |
|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20110329 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20110405 |
|
| A521 | Request for written amendment filed |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20110525 |
|
| TRDD | Decision of grant or rejection written | ||
| A01 | Written decision to grant a patent or to grant a registration (utility model) |
Free format text: JAPANESE INTERMEDIATE CODE: A01 Effective date: 20110614 |
|
| A01 | Written decision to grant a patent or to grant a registration (utility model) |
Free format text: JAPANESE INTERMEDIATE CODE: A01 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20110616 |
|
| R150 | Certificate of patent or registration of utility model |
Ref document number: 4768499 Country of ref document: JP Free format text: JAPANESE INTERMEDIATE CODE: R150 Free format text: JAPANESE INTERMEDIATE CODE: R150 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20140624 Year of fee payment: 3 |
|
| LAPS | Cancellation because of no payment of annual fees |