Camera and radar fusion
A system includes a computer that includes a processor and a memory. The memory stores instructions executable by the processor such that the computer is programmed to transform a vertical image plane camera image to a horizontal image plane, thereby generating a transformed camera image, generate a scaled horizontal image plane radar image based on a resolution of the transformed camera image, and then superimpose the scaled radar image on the transformed camera image thereby generating a horizontal image plane superimposed image.
1. A system, comprising a computer that includes a processor and a memory, the memory storing instructions executable by the processor such that the computer is programmed to:
transform a vertical image plane camera image to a horizontal image plane, thereby generating a transformed camera image;
generate a scaled horizontal image plane radar image based on a resolution of the transformed camera image;
then superimpose the scaled radar image on the transformed camera image thereby generating a horizontal image plane superimposed image; and
determine an inverse transform function that transforms the horizontal image plane superimposed image to the vertical image plane, thereby generating a vertical image plane superimposed image that includes longitudinal coordinates of objects.
2. The system of claim 1 , wherein the instructions further include instructions to generate the scaled radar image further based on an occupancy grid map image extracted from received radar image.
3. The system of claim 1 , wherein the instructions further include instructions to:
detect an object in the camera image;
identify a first location of the detected object in the transformed camera image;
identify a second location of the detected object in the radar image; and
rescale the scaled radar image to superimpose the first location of the object in the radar image on the second location of the object in the transformed camera image.
4. The system of claim 1 , wherein the instructions to generate the vertical image plane superimposed image further include:
applying the inverse transform function to the radar image thereby generating a vertical image plane radar image; and
superimposing the vertical image plane radar image on the vertical plane camera image.
5. The system of claim 1 , wherein the instructions further include instructions to:
detect, in the transformed camera image, two or more road markings;
determine a number of pixels between the detected two or more road markings in the transformed image; and
determine the resolution of the camera image based on a real-world distance of the detected two or more road markings and the determined number of pixels between the detected road markings in the transformed image.
6. The system of claim 1 , wherein the instructions further include instructions to crop the radar image based on an overlapping of fields of view of the camera and image data and generate the scaled radar image from the cropped radar image.
7. The system of claim 1 , wherein the vertical image plane camera image is distorted and the instructions further include instructions to undistort the camera image prior to transforming the camera image to the horizontal image plane.
8. The system of claim 5 , wherein the instructions further include instructions to determine the resolution of the camera image based on real-world dimensions of at least one of a guard rail, light pole, traffic sign, and wall, detected in the transformed camera image.
9. A system, comprising:
a radar sensor with a first field of view and a camera sensor with a second field of view, wherein the first and second fields of view at least partially overlap; and
a computer programmed to:
transform a vertical image plane camera image to a horizontal image plane, thereby generating a transformed camera image;
generate a scaled horizontal image plane radar image based on a resolution of the transformed camera image;
then superimpose the scaled radar image on the transformed camera image thereby generating a horizontal image plane superimposed image; and
determine an inverse transform function that transforms the horizontal image plane superimposed image to the vertical image plane, thereby generating a vertical image plane superimposed image that includes longitudinal coordinates of objects.
10. A method, comprising:
transforming a vertical image plane camera image to a horizontal image plane, thereby generating a transformed camera image;
generating a scaled horizontal image plane radar image based on a resolution of the transformed camera image;
then superimposing the scaled radar image on the transformed camera image thereby generating a horizontal image plane superimposed image; and
determining an inverse transform function that transforms the horizontal image plane superimposed image to the vertical image plane, thereby generating a vertical image plane superimposed image that includes longitudinal coordinates of objects.
11. The method of claim 10 , further comprising generating the scaled radar image further based on an occupancy grid map image extracted from received radar image.
12. The method of claim 10 , further comprising:
detecting an object in the camera image;
identifying a first location of the detected object in the transformed camera image;
identifying a second location of the detected object in the radar image;
and rescaling the scaled radar image to superimpose the first location of the object in the radar image on the second location of the object in the transformed camera image.
13. The method of claim 10 , wherein generating the vertical image plane superimposed image further includes:
applying the inverse transform function to the radar image thereby generating a vertical image plane radar image; and
superimposing the vertical image plane radar image on the vertical plane camera image.
14. The method of claim 10 , further comprising:
detecting, in the transformed camera image, two or more road markings;
determining a number of pixels between the detected two or more road markings in the transformed image; and
determining the resolution of the camera image based on a real-world distance of the detected road markings and the determined number of pixels between the detected two or more road markings in the transformed image.
15. The method of claim 10 , further comprising cropping the radar image based on an overlapping of fields of view of the camera and image data and generating the scaled radar image from the cropped radar image.
16. The method of claim 10 , further comprising undistorting the vertical image plane camera image data prior to transforming the camera image to the horizontal image plane, wherein the received camera image is distorted.
17. The method of claim 14 , further comprising determining the resolution of the camera image based on real-world dimensions of at least one of a guard rail, light pole, traffic sign, and wall, detected in the transformed camera image.