OBSTACLE DETECTION DEVICE AND METHOD AND OBSTACLE DETECTION SYSTEM
First Claim
1. An obstacle detection device for detecting an obstacle around a vehicle from a spectral image obtained by an imaging unit mounted on the vehicle, the obstacle detection device comprising:
- an image storage unit for storing an image provided by the imaging unit;
a flow extraction unit for extracting a flow of a pixel from images stored at two different times in the image storage unit, the two different times being a current time and an earlier time than the current time;
a vehicle movement amount estimation unit for estimating an amount of movement of the vehicle in a real world based on the extracted flow;
a road plane estimation unit for estimating a position of a road plane based on the extracted flow and the amount of movement of the vehicle;
an obstacle candidate point relating unit relating pixels in images at the two times by using a pixel in an image at an earlier time as a reference, by assuming that the pixel corresponds to the road plane in the real world, and by extracting a corresponding pixel at the current time based on the amount of movement of the vehicle;
a non-matching range identifying unit for comparing, pixel by pixel, partial images having centers at obstacle candidate points related between the pixels in the images at the two times, and identifying a range in which pixel values are different as a non-matching range; and
an obstacle determining unit for comparing the pixel values in the spectral images at the two times corresponding to the non-matching range, and determining whether the obstacle candidate point comprises a point corresponding to a shadow on the road plane or a point corresponding to an actual obstacle.
1 Assignment
0 Petitions
Accused Products
Abstract
An obstacle region candidate point relating unit assumes that a pixel in an image corresponds to a point on a road surface, and associates pixels between images at two times on the basis of the amount of movement of a vehicle in question, a road plane, and a flow of the image estimated. When a pixel corresponds to a shadow of the vehicle in question or the moving object therearound appearing on the road surface, the ratio of intensities of the pixel values of the spectral images between two images should be approximately the same as the ratio of the spectral characteristics of the sunshine in the sun and the shade. Therefore, when the ratio of intensities is approximately the same as the ratio of the spectral characteristics, the obstacle determining unit does not determine that the pixel in question is a point corresponding to the obstacle. Only when the ratio of intensities is not approximately the same as the ratio of the spectral characteristics, the obstacle determining unit determines that the pixel in question is a point corresponding to the obstacle.
40 Citations
17 Claims
-
1. An obstacle detection device for detecting an obstacle around a vehicle from a spectral image obtained by an imaging unit mounted on the vehicle, the obstacle detection device comprising:
-
an image storage unit for storing an image provided by the imaging unit; a flow extraction unit for extracting a flow of a pixel from images stored at two different times in the image storage unit, the two different times being a current time and an earlier time than the current time; a vehicle movement amount estimation unit for estimating an amount of movement of the vehicle in a real world based on the extracted flow; a road plane estimation unit for estimating a position of a road plane based on the extracted flow and the amount of movement of the vehicle; an obstacle candidate point relating unit relating pixels in images at the two times by using a pixel in an image at an earlier time as a reference, by assuming that the pixel corresponds to the road plane in the real world, and by extracting a corresponding pixel at the current time based on the amount of movement of the vehicle; a non-matching range identifying unit for comparing, pixel by pixel, partial images having centers at obstacle candidate points related between the pixels in the images at the two times, and identifying a range in which pixel values are different as a non-matching range; and an obstacle determining unit for comparing the pixel values in the spectral images at the two times corresponding to the non-matching range, and determining whether the obstacle candidate point comprises a point corresponding to a shadow on the road plane or a point corresponding to an actual obstacle. - View Dependent Claims (2, 3, 4, 9, 11, 12, 15, 16, 17)
-
-
5. An obstacle detection method for detecting an obstacle around a vehicle from a spectral image obtained by an imaging unit mounted on the vehicle, the obstacle detection method comprising:
-
causing an image storage unit to store a image provided by the imaging unit; extracting a flow of a pixel from images stored at two different times in the image storage unit, the two different times being a current time and an earlier time than the current time; estimating an amount of movement of the vehicle in a real world based on the extracted flow; estimating a position of road plane based on the extracted flow and the amount of movement of the vehicle; relating pixels in images at the two times by using a pixel in an image at an earlier time as a reference, assuming that the pixel corresponds to the road plane in the real world, and extracting a corresponding pixel at the current time based on the amount of movement of the vehicle; comparing, pixel by pixel, partial images having centers at obstacle candidate points related between the pixels in the images at the two times, and identifying a range in which pixel values are different as a non-matching range; and comparing the pixel values in the spectral images at the two times corresponding to the non-matching range, and determining whether the obstacle candidate point comprises a point corresponding to a shadow on the road plane or a point corresponding to an actual obstacle. - View Dependent Claims (6, 7, 8, 13, 14)
-
-
10. A computer-readable information recording medium storing a program for causing a computer to function as a flow extraction unit for extracting a flow of a pixel from images stored at two different times in an image storage unit, the two different times being a current time and an earlier time than the current time;
-
a vehicle movement amount estimation unit for estimating an amount of movement of the vehicle in a real world based on the extracted flow; a road plane estimation unit for estimating a position of road plane based on the extracted flow and the amount of movement of the vehicle; an obstacle candidate point relating unit relating pixels in images at the two times by using a pixel in an image at an earlier time as a reference, assuming that the pixel corresponds to the road plane in the real world, and extracting a corresponding pixel at the current time based on the amount of movement of the vehicle; a non-matching range identifying unit for comparing, pixel by pixel, partial images having centers at obstacle candidate points relater between the pixels in the images at the two times, and identifying a range in which pixel values are different as a non-matching range; and an obstacle determining unit for comparing the pixel values in the spectral images at the two times corresponding to the non-matching range, and determining whether the obstacle candidate point comprises a point corresponding to a shadow on the road plane or a point corresponding to an actual obstacle.
-
Specification