Vehicle detection apparatus and method

• US 7,522,747 B2
• Filed: 02/01/2005
• Issued: 04/21/2009
• Est. Priority Date: 03/26/2004
• Status: Expired due to Fees

• Current Assignee: Omron Automotive Electronics Company Limited (Nidec Corporation)
• Original Assignee: Omron Corporation
• Inventors: Horibe, Koji
• Primary Examiner(s): Bella; Matthew C
• Assistant Examiner(s): Seth; Manav
• Application Number:
  US11/050,043
• Publication Number:
  US 20050213794A1
• Time in Patent Office: 1,540 Days
• Field of Search: 382103-107, 382/100, 382/209, 382217-222, 382305-306, 342/70, 342/118, 342/357.01
• US Class Current: 382/104
• CPC Class Codes: B60R 2300/301 combining image information with other obstacle sensor information, e.g. using RADAR/LIDAR/SONAR sensors for estimating risk of collision
• CPC Class Codes: B60R 2300/8093 for obstacle warning
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

A front-going vehicle is recognized by using a camera image from a camera for monitoring the front of one'"'"'s own vehicle and detection data from a radar for detecting position of an object in front of one'"'"'s own vehicle and distance to this object. A front-going vehicle is recognized among photographed objects with object images contained in the camera image by extracting images in a specified area inside the camera image centered around positions of objects in detection data from the radar, binarizing the extracted images, matching patterns between an image figure contained in the binarized image and reference images and judging an image figure as a figure of a front-going vehicle if it is concluded that this image figure matches one of the reference images over a specified percentage. The reference figures are of a specified pattern, having a horizontally elongated part with downward protrusions from both its end parts so as to resemble a back view of a vehicle.

Claims

1. A vehicle detection apparatus for using a camera image from a camera for monitoring the front of one'"'"'s own vehicle and detection data from a radar for detecting position of an object in front of said own vehicle and distance to said object from said own vehicle to recognize a front-going vehicle among photographed objects;

said vehicle detection apparatus comprising;

a memory;

registering means for preliminarily registering in said memory reference images each with a size corresponding to the distance to said object, each of said reference images having a horizontally elongated part with downward protrusions that protrude only downward from both end parts of said horizontally elongated part and having no upward protrusions that protrude upward from said horizontally elongated part;

image extracting means for extracting images in a specified area inside said camera image centered around positions of objects in the detection data from said radar;

binarization means for binarizing the extracted images by said image extracting means;

matching means for matching patterns between an image figure contained in the binarized image and said reference images to look up one of said reference images with a size matching a detected distance; and

judging means for judging an image figure as a figure of a front-going vehicle if said matching means concludes that said image figure matches one of said reference images over a specified percentage.

3. The vehicle detection apparatus of claim 1 wherein said specified area is smaller than said camera image.

2. A method of using a camera image from a camera for monitoring the front of one'"'"'s own vehicle and detection data from a radar for detecting position of an object in front of said own vehicle and distance to said object from said own vehicle to recognize a front-going vehicle among photographed objects, said method comprising the steps of:

using a processor to perform the steps of;

preliminarily registering in a memory a plurality of reference images each with a size corresponding to the distance to said object, each of said reference images having a horizontally elongated part with downward protrusions that protrude only downward from both end parts of said horizontally elongated part and having no upward protrusions that protrude upward from said horizontally elongated part;

extracting images in a specified area inside said camera image centered around positions of objects in the detection data from said radar;

binarizing the extracted images;

matching patterns between an image figure contained in the binarized image and reference images to look up one of said reference images with a size matching a detected distance; and

judging an image figure as a figure of a front-going vehicle if it is concluded by the step of matching that said image figure matches one of said reference images over a specified percentage.

4. The method of claim 2 wherein said specified area is smaller than said camera image.