A method for generating surround view images derived from image data captured by cameras of a vehicular surround view vision system includes equipping a vehicle with a plurality of cameras disposed at the vehicle. Image data is captured by the cameras and provided to a control of the vehicle. The provided captured image data is processed to generate a first three-dimensional representation in accordance with a first curved surface model as if seen by a virtual observer from a first virtual viewing point exterior of the vehicle having a first viewing direction. The first representation is output to a display screen of the vehicle for display to the driver of the vehicle. Responsive to actuation of a user input, the processing is adjusted to generate a second three-dimensional representation in accordance with a second curved surface model, and the second representation is output to the display screen.