Aiming at solving the periodic defects of nozzle clogging and step error in printing process

the method of digital printing detect detection based on wavelet decomposition and local binary pattern variance（ LBPV）feature is proposed. Firstly

the printing defect image is decomposed by wavelet transform and fused with specific high-frequency detail sub-images to effectively suppress the background texture of the printing image. Then

the local binary pattern variance feature of the fused sub-image is calculated to reduce the dimension of the low-texture feature. Finally

the feature map is projected in horizontal direction and the number of pixels of specific LBPV value is calculated. Furthermore

the digital printing defects was detected and located by the peak value of the projection curve. In our research

the proposed algorithm can detect the printing and realize the visual recognition of defect locations accurately.