Traditional bridge crack detection methods are low accuracy and low ef ficiency.A new method of bridge crack detection based on image regeneration was proposed.First,real bridge crack image that collected by unmanned aerial vehi cle was subtracted from the same size pure white image.Second,image enhancement was carried out based on the gamma transformation algorithm.Third,removed the background,retained cracks of the image and regenerated the crack image with bl ack background by using the histogram of image.Finally,extracted and located c racks by using contour search function.Before finding the contour,a kernel is setted for fuzzy denoising,and the iterator is used to remove the noise contour .Finally,in order to ensure the continuity and integrity of the crack,the wat er-filling is performed by calculating the centroid coordinates of the contour w hen drawing the contour.Experimental results show that the crack accuracy index and crack recall index are up to 98%,and the qualified rate of crack location can reach 92.4% when the location accuracy S<30pixels.