Abstract: To address the lack of effective quality detection methods for cigarette online laser perforation, a measurement method for cigarette online laser perforation parameters was proposed based on computer vision technology. Firstly, a panoramic image of the filter perforated area was captured and spliced using a rotation imaging device for cigarettes. The image contrast of the perforated area was effectively highlighted through the lighting system at the filter end. Then, a self-adaptive segmentation algorithm based on directional projection was used to achieve accurate segmentation and extraction of small holes. Finally, the detections of distance between hole rows, pitch of holes and hole area were achieved by measuring the morphological parameters of the target area. The results showed that: 1) The image segmentation and contour extraction using the proposed small hole region segmentation algorithm based on directional projection both had a high peak signal-to-noise ratio(PSNR)and a structural similarity(SSIM) index close to 1, showing a high degree of consistence with the true results and significantly better segmentation performance than the universal threshold segmentation method. 2) The average relative deviations of the 10 repeated measurements of the distance between hole rows, hole area and pitch of holes for the measuring samples with standard perforations were 0.30%, 0.72% and 0.41%, respectively, indicating high measurement accuracy of the proposed method. 3)The maximum variable coefficients for the distance between hole rows, pitch of holes and hole area of actual online laser perforated cigarette samples were 0.54%, 0.01% and 0.95%, respectively, indicating good repeatability of the proposed method. The small hole region segmentation algorithm based on the directional projection proposed in this study is capable of obtaining accurate perforation parameter information of perforated cigarettes, providing technical support for the correlation between perforation parameters and parameters such as cigarette ventilation.