US 11,889,047 B2
Image processing device and image processing method
Minoru Kuraoka, Tokyo (JP)
Assigned to SUBARU CORPORATION, Tokyo (JP)
Filed by SUBARU CORPORATION, Tokyo (JP)
Filed on Apr. 1, 2021, as Appl. No. 17/220,095.
Claims priority of application No. 2020-067353 (JP), filed on Apr. 3, 2020.
Prior Publication US 2021/0312159 A1, Oct. 7, 2021
Int. Cl. G06T 7/593 (2017.01); H04N 13/239 (2018.01); G06V 20/64 (2022.01); G06V 20/56 (2022.01); G06V 10/25 (2022.01); G06V 10/75 (2022.01)
CPC H04N 13/239 (2018.05) [G06T 7/593 (2017.01); G06V 10/25 (2022.01); G06V 10/751 (2022.01); G06V 20/56 (2022.01); G06V 20/64 (2022.01); G06T 2207/10012 (2013.01); G06T 2207/30252 (2013.01); G06V 2201/08 (2022.01)] 19 Claims
OG exemplary drawing
 
1. An image processing device comprising
one or more processors communicably connected to a stereo camera, the one or more processors being configured to:
receive a stereo image from the stereo camera;
set a three-dimensional object region by detecting a three-dimensional object, on a basis of a distance image that is to be generated on a basis of the stereo image and includes parallax-related values corresponding to parallaxes in respective pixels;
set, into the distance image, a first image region that is same as the three-dimensional object region in size and position, a second image region including a left end of the first image region and partly overlapping the first image region, and a third image region including a right end of the first image region and partly overlapping the first image region;
calculate, in each of the first image region, the second image region, and the third image region, representative values of the parallax-related values including (i) a first representative value based on first parallax-related values corresponding to a first pixel column that includes pixels having a first horizontal coordinate in the distance image, and (ii) a second representative value based on second parallax-related values corresponding to a second pixel column that includes pixels having a second horizontal coordinate different from the first horizontal coordinate in the distance image,
calculate a first approximate line in the first image region based on the representative values in the first image region, a second approximate line in the second image region based on the representative values in the second image region, and a third approximate line in the third image region based on the representative values in the third image region;
estimate a continuous structure degree indicating a degree of likeliness of the three-dimensional object being a continuous structure, on a basis of a first slope value of the first approximate line of the first image region, a second slope value of the second approximate line of the second image region, and a third slope value of the third approximate line of the third image region; and
determine whether the three-dimensional object detected by the detector is a vehicle, on a basis of the continuous structure degree.