Vehicle environment recognition system

Claim 2

A vehicle driving environment recognition system comprising: imaging means for periodically forming an input image of a forward road ahead of a road vehicle; lateral position detecting means for detecting a lateral position of the vehicle relative to a road in accordance with the input image; image integrating means for transforming the input image into a transformed image in a road plane coordinate system by using the lateral position; and for producing an integral image by integrating the transformed image; and feature extracting means for extracting a characteristic region from the integral image.

Claim 3

The recognition system as claimed in claim 1 wherein the transforming means includes means for determining equations of transformation for transforming the input image to the road plane coordinate system in accordance with parameters of the imaging means, and for transforming the input image into the transformed image in the road plane coordinate system by using the lateral position and the equations.

Claim 4

The recognition system as claimed in claim 1 wherein the image integrating means includes means for calculating a lateral resolution at each pixel location in accordance with the parameters of the imaging means and for transforming the input image into the transformed image by shifting each pixel location by an amount determined in accordance with the lateral resolution and a lateral displacement of the vehicle determined from the lateral position detected by the lateral position detecting means.

Claim 5

The recognition system as claimed in claim 1 wherein the lateral position detecting means comprises white line detecting means for detecting a road white line from the input image and for calculating the lateral position of the vehicle relative to the road from a position of the white line.

Claim 6

The recognition system as claimed in claim 1 wherein the lateral position detecting means comprises means for detecting vehicle guide means provided in a road, and calculating the lateral position of the vehicle relative to the road from a position of the vehicle guide means.

Claim 7

The recognition system as claimed in claim 1 wherein the feature extracting means includes a means for detecting a position of a preceding vehicle from the integral image.

Claim 8

A vehicle driving environment recognition system comprising: imaging means for forming an input image of a forward road ahead of a road vehicle; lateral position detecting means for detecting a lateral position of the vehicle relative to a road; image transforming means for determining an image transforming method in accordance with a characteristic parameter and a positional parameter of the imaging means and transforming the input image into a transformed image by using the lateral position; integrating means for producing an integral image by integrating the transformed image; and feature extracting means for extracting a characteristic region from the integral image and for detecting a shape of the forward road from the integral image.

Claim 9

A vehicle driving environment recognition system comprising: imaging means for forming an input image of a forward road ahead of a road vehicle; lateral position detecting means for detecting a lateral position of the vehicle relative to a road; image transforming means for determining an image transforming method in accordance with a characteristic parameter and a positional parameter of the imaging means and transforming the input image into a transformed image by using the lateral position; integrating means for producing an integral image by integrating the transformed image; and feature extracting means for extracting a characteristic region from the integral image and for detecting a shape of the forward road from the integral image.

Claim 10

A vehicle environment recognition system comprising: an input section for collecting information on a driving environment of a road vehicle, the input section forming an input image of a forward scene ahead of the vehicle and supplying data to determine a lateral position of the vehicle relative to a road and data to recognize the environment; and a processing section for producing a modified image by modifying the input image by a coordinate transformation from an image plane coordinate system to a road plane coordinate system in accordance with the lateral position of the vehicle, further producing an integral image by using the modified image, and for extracting an image feature from the integral image.

Claim 11

A vehicle environment recognition system comprising: an input section for collecting information on a driving environment of a road vehicle, the input section forming an input image of a forward scene ahead of the vehicle and supplying data to determine a lateral position of the vehicle relative to a road and data to recognize the environment; and a processing section for producing a modified image by modifying the input image by a coordinate transformation from an image plane coordinate system to a road plane coordinate system in accordance with the lateral position of the vehicle, further producing an integral image by using the modified image, and for extracting an image feature from the integral image.

Claim 12

The recognition system as claimed in claim 8 wherein the input section comprises a camera for producing a time series of input images of a scene ahead of the vehicle, the processing section is configured to produce a time series of integral images in response to the time series of input images supplied from the camera, the processing section is configured to produce a new integral image from a current image prepared from a most recent input image supplied from the camera and a modified previous image which is an image transformed from a most recent integral image in accordance with the lateral position.

Claim 13

The recognition system as claimed in claim 9 wherein the processing section is configured to determine a lateral vehicle displacement from a change in the lateral position, and to produce the modified previous image by displacing each pixel location along a horizontal axis of a coordinate system of the most recent integral image by an amount of lateral image displacement determined in accordance with the lateral vehicle displacement.

Claim 14

A recognition system comprising: an input section for collecting information on a driving environment of a road vehicle, the input section forming an input image of a forward scene ahead of the vehicle and supplying data to determine a lateral position of the vehicle relative to a road and data to recognize the environment; and a processing section for producing a modified image by modifying the input image in accordance with the lateral position of the vehicle, further producing an integral image by using the modified image, and for extracting an image feature from the integral image, wherein the input section comprises a camera for producing a time series of input images of a scene ahead of the vehicle, and the processing section is configured to produce a time series of integral images in response to the time series of input images supplied from the camera, produce a new integral image from a current image prepared from a most recent input image supplied from a camera and a modified previous image which is an image transformed from a most recent integral image in accordance with the lateral position, determine a lateral vehicle displacement from a change in the lateral position, produce the modified previous image by displacing each pixel location along to horizontal axis of a coordinate system of the most recent integral image by an amount of lateral image displacement determined in accordance with the lateral vehicle displacement, and produce the new integral image by determining a weighted average of image values of the current image and the modified previous image in each elemental image region.

Claim 15

A recognition system comprising: an input section for collecting information on a driving environment of a road vehicle, the input section forming an input image of a forward scene ahead of the vehicle and supplying data to determine a lateral position of the vehicle relative to a road and data to recognize the environment; and a processing section for producing a modified image by modifying the input image in accordance with the lateral position of the vehicle, further producing an integral image by using the modified image, and for extracting an image feature from the integral image, wherein the input section comprises a camera for producing a time series of input images of a scene ahead of the vehicle, and the processing section is configured to produce a time series of integral images in response to the time series of input images supplied from the camera, produce a new integral image from a current image prepared from a most recent input image supplied from a camera and a modified previous image which is an image transformed from a most recent integral image in accordance with the lateral position, determine a lateral vehicle displacement from a change in the lateral position, produce the modified previous image by displacing each pixel location along to horizontal axis of a coordinate system of the most recent integral image by an amount of lateral image displacement determined in accordance with the lateral vehicle displacement, and produce the new integral image by determining a weighted average of image values of the current image and the modified previous image in each elemental image region.

Claim 16

The recognition system as claimed in claim 11 wherein the processing section is configured to produce a transformed image from each input image by a coordinate transformation from an image plane coordinate system to a road plane coordinate system determined by camera parameters, to prepare the modified previous image by translating the most recent integral image along a horizontal axis of the road coordinate system in a direction opposite to a direction of the lateral vehicle displacement by the amount of the lateral image displacement, and to produce the new integral image by superposing the modified previous image on the current image which is a most recent transformed image.

Claim 17

The recognition system as claimed in claim 12 wherein the processing section is configured to produce a transformed image from each input image according to equations of coordinate transformation determined in accordance with the camera parameters which comprise a height of the camera, a depression angle of the camera and a focal length of the camera.

Claim 18

The recognition system as claimed in claim 11 wherein the processing section is configured to shift each pixel location in the most recent integral image by the amount of the lateral image displacement determined in accordance with the lateral vehicle displacement and a lateral resolution, the lateral resolution being determined for each pixel location in accordance with camera parameters which comprise a height of the camera, a depression angle of the camera and a focal length of the camera.

Claim 19

The recognition system as claimed in claim 10 wherein the processing section is configured to detect a road lane boundary in each input image and to determine the lateral position of the vehicle relative to the road from a position of the road lane boundary in each input image.

Claim 20

The recognition system as claimed in claim 10 wherein the input section further comprises a device for sensing a signal from a road, and the processing section is configured to determined the lateral position of the vehicle relative to the road in accordance with the signal.

Claim 21

The recognition system as claimed in claim 10 wherein the processing section is configured to detect one of a position of a preceding vehicle ahead of the road vehicle in which the recognition system is installed, and a configuration of a forward road.

Claim 22

The recognition system as claimed in claim 1 wherein the image integrating means includes means for integrating the transformed image by shifting one of a current transformed image and a previous transformed image in accordance with a different between the lateral position in the previous transformed image and the lateral position in the current transformed image.