The invention discloses a lane departure early warning method based on inverse perspective transformation. The method comprises the following steps that road images are collected, image preprocessingis performed on the collected road image information, and vanishing point information and lane line information are extracted from the preprocessed images; yaw angle gamma is obtained according to thevanishing point information; pixel coordinate conversion is carried out on the preprocessed road image to obtain an inverse perspective transformation top view, and vehicle lateral deviation d6 in the actual road top view is obtained through a distance coordinate conversion algorithm by the lane line information, the yaw angle gamma and the inverse perspective transformation top view; and cross-track time TLC and a cross-track time change rate eta are calculated according to the yaw angle gamma and the lateral deviation D6. According to the lane departure early warning method based on the inverse perspective transformation, the lane departure safety risk assessment and early warning are carried out according to the TLC and the cross-track time change rate eta, and compared with a traditional deviation early-warning algorithm based on cross-track time, the lane departure safety risk assessment and early warning are more accurate and reasonable.