Image processing apparatus and image transformation method

1. An image processing apparatus comprising:

an image acquisition unit that acquires a captured image with a first spatial resolution;

a distance acquisition unit that acquires a distance image which is depth information with a second spatial resolution that is a resolution lower than the first spatial resolution;

an image recognition unit that extracts an area including a three-dimensional object area corresponding to a three-dimensional object in the captured image;

a distance calculation unit that calculates depth information of the three-dimensional object area on the basis of the distance image;

a correction unit that corrects coordinate transformation information for coordinate transformation of the captured image on the basis of the depth information of the three-dimensional object area calculated by the distance calculation unit; and

a viewpoint transformed image generation unit that generates a viewpoint transformed image obtained by the coordinate transformation of the captured image by using the coordinate transformation information corrected by the correction unit.

2. The image processing apparatus according to claim 1 ,

wherein the image recognition unit extracts outlines of a plurality of areas including the three-dimensional object area on the basis of at least one of luminance, hue, color saturation, and brightness of the captured image.

3. The image processing apparatus according to claim 1 ,

wherein the coordinate transformation information includes a plurality of combinations between transformation source coordinates in the captured image and transformation destination coordinates in the viewpoint transformed image; and

wherein the correction unit corrects the transformation destination coordinates on the basis of the depth information of the three-dimensional object area.

4. The image processing apparatus according to claim 3 ,

wherein the coordinate transformation information is created in advance on a premise that an object in the captured image is an area on a road surface;

wherein the image recognition unit divides the captured image into a plurality of areas with respect to respective objects included in the captured image; and

wherein the correction unit judges whether each of the plurality of areas is the three-dimensional object area or not, based on the distance image and corrects the transformation destination coordinates corresponding to the transformation source coordinates in an area which is judged as the three-dimensional object area.

5. The image processing apparatus according to claim 1 ,

wherein the coordinate transformation information includes a plurality of combinations between transformation source coordinates in the captured image and transformation destination coordinates in the viewpoint transformed image; and

wherein the correction unit corrects the transformation source coordinates to be located in the vicinity of an outline of the three-dimensional object area.

6. The image processing apparatus according to claim 5 ,

wherein the correction unit locates the transformation source coordinates in the vicinity of the outline of the three-dimensional object area and outside and inside the three-dimensional object area.

7. The image processing apparatus according to claim 1 ,

wherein the coordinate transformation information is calculated on a premise that an object in the captured image exists within a reference height plane;

wherein the correction unit calculates a height of the object from the reference height plane and a position of the object on the reference height plane by using the depth information of the three-dimensional object area which is calculated by the distance calculation unit; and

wherein the correction unit corrects the coordinate transformation information by using a viewpoint transformation reference position, the calculated height of the object from the reference height plane, and the calculated position of the object on the reference height plane.

8. An image transformation method comprising:

acquiring a captured image with a first spatial resolution;

acquiring a distance image which is depth information with a second spatial resolution that is a resolution lower than the first spatial resolution;

extracting an area including a three-dimensional object area corresponding to a three-dimensional object in the captured image;

calculating depth information of the three-dimensional object area on the basis of the distance image;

correcting coordinate transformation information for coordinate transformation of the captured image on the basis of the calculated depth information of the three-dimensional object area; and

generating a viewpoint transformed image obtained by the coordinate transformation of the captured image by using the coordinate transformation information corrected by the correction.

9. The image transformation method according to claim 8 ,

wherein the extraction of the area including the three-dimensional object area is executed on the basis of at least one of luminance, hue, color saturation, and brightness of the captured image.

10. The image transformation method according to claim 8 ,

wherein the coordinate transformation information includes a plurality of combinations between transformation source coordinates in the captured image and transformation destination coordinates in the viewpoint transformed image; and

wherein the transformation destination coordinates are corrected on the basis of the depth information of the three-dimensional object area.

11. The image transformation method according to claim 10 ,

wherein the coordinate transformation information is created in advance on a premise that an object in the captured image is an area on a road surface;

wherein the captured image is divided into a plurality of areas with respect to respective objects included in the captured image; and

wherein whether each of the plurality of areas is the three-dimensional object area or not is judged based on the distance image and the transformation destination coordinates corresponding to the transformation source coordinates in an area which is judged as the three-dimensional object area are corrected.

12. The image transformation method according to claim 8 ,

wherein the coordinate transformation information includes a plurality of combinations between transformation source coordinates in the captured image and transformation destination coordinates in the viewpoint transformed image; and

wherein the correction based on the depth information of the three-dimensional object area includes correction of the transformation source coordinates to be located in the vicinity of an outline of the three-dimensional object area.

13. The image transformation method according to claim 12 ,

wherein the correction based on the depth information of the three-dimensional object area includes locating the transformation source coordinates in the vicinity of the outline of the three-dimensional object area and outside and inside the three-dimensional object area.

14. The image transformation method according to claim 8 ,

wherein the coordinate transformation information is calculated on a premise that an object in the captured image exists within a reference height plane; and

wherein the correction of the coordinate transformation information calculates a height of the object from the reference height plane and a position of the object on the reference height plane by using the calculated depth information of the three-dimensional object area and is performed by using a viewpoint transformation reference position, the calculated height of the object from the reference height plane, and the calculated position of the object on the reference height plane.