The RALPH self driving Navlab 5

Possible use as a web based teaching example.

https://docs.opencv.org/4.x/dd/d52/tutorial_js_geometric_transformations.html Fast transformation of the road view, see Perspective Transformation.

let src = cv.imread('canvasInput');
let dst = new cv.Mat();
let dsize = new cv.Size(src.rows, src.cols);

let x_cut_left = Math.round(src.cols*0.01) //distance in from the left where the near road starts
let x_cut_right = 0//Math.round(src.cols*0.01) //what to cut off the right
let road_center = 0.74 //percentage of center position of the road
let x_center = Math.round((src.cols - x_cut_left - x_cut_right)*road_center)
let x_road_far = Math.round(src.cols * 0.2) //width of the road in the distance

let y_road_far = Math.round(src.rows*.45) //cut off the skyline
let y_road_near = Math.round(src.rows *0.65) //cut off the car hood and dash

let srcTri = cv.matFromArray(4, 1, cv.CV_32FC2, [ //upperleft, upperright, lowerleft, lowerright
    x_cut_left+x_center-x_road_far, y_road_far, 
    x_center+x_road_far-x_cut_right, y_road_far, 
    x_cut_left, y_road_near, 
    src.cols-x_cut_right, y_road_near
    ]);
let dstTri = cv.matFromArray(4, 1, cv.CV_32FC2, [0, 0, src.cols, 0, 0, src.rows, src.cols, src.rows]);
let M = cv.getPerspectiveTransform(srcTri, dstTri);
// You can try more different parameters
cv.warpPerspective(src, dst, M, dsize, cv.INTER_LINEAR, cv.BORDER_CONSTANT, new cv.Scalar());
cv.imshow('canvasOutput', dst);
src.delete(); dst.delete(); M.delete(); srcTri.delete(); dstTri.delete();

https://app.qoom.io/tutorials/DrivingSimulator/guide.md A driving simulator could be used to provide input and test output.

https://www.kaggle.com/datasets/roydatascience/training-car Datasets of training data are availale with csv data indicating the degree of turn.