Abstract: Aiming at the problem that it is difficult to identify pavement cracks from complex pavement backgrounds effectively，a method of pavement crack image recognition based on Fractional Fourier Transform（FrFT）and spectrum enhancement processing was proposed. Firstly，the image was pre-processed using the blocking method and median filtering；secondly，the fracture image was converted to a spectrogram by a FrFT，and the spectrum was subjected to region removal and enhancement processing. Fractional inverse Fourier transform was used to convert the processed spectrum image into an enhanced crack image；then，the defect template was selected by image superposition，inversion and threshold filtering on the enhanced crack image；finally，Canny edge detection was used in combination with expansion corrosion and area screening threshold method to identify pavement cracks based on defect templates. Compared with the Wavelet Transform （WT） detection algorithm，the detection method based on FrFT has 14. 35% and 17. 1% improvements in accuracy and recall，respectively，and the influence of the spectral enhancement multiple and the minus threshold on the experimental results was analyzed. Experimental results show that the method of Fractional Fourier Transform and spectral enhancement processing can effectively improve the recognition rate of pavement cracks.