The invention discloses a zebra crossing real-time detection method based in intelligent driving, and belongs to the field of traffic information detection in the intelligent traffic industry. The zebra crossing real-time detection method includes the steps of firstly, obtaining an image Src_Image of the road in front of an intelligent vehicle in real time at a frame rate of 25 frames per second to 50 frames per second, and conducting inverse perspective transformation on the obtained original image Src_Image to obtain a bird view image Bird_View of a road image; then, cutting a lane area image ROI_Image out of the bird view image Bird_View according to two lane lines of a lane, conducting graying, self-adaption binaryzation, canny edge extraction and morphology corrosion expansion on the image ROI_Image to obtain a binaryzation image Dst_Bw; finally, conducting transverse pixel statistics on the image Dst_Bw to obtain the black-white skip time number Sum, the black strip average width Wb, the white strip average width Ww and the strip average height H. If the conditions that Sum is larger than S1 and smaller than S2, the absolute value of the difference between Wb and Ww is smaller than W, Wb is larger than Wb1 and smaller than Wb2, Ww is larger than Ww1 and smaller than Ww2, and H is larger than H1 and smaller than H2 are all satisfied, the zebra crossing is confirmed to appear in front of the intelligent vehicle, and if not, the zebra crossing is confirmed not to appear.