The complex internal environment of urban drainage pipelines seriously affects the accurate location of internal defects. Therefore, research on defect localization technology is carried out for urban drainage pipelines. This study introduces an improved Canny algorithm to estimate the internal deviation distance of urban drainage pipelines. By combining Histogram of Oriented Gradients (HOG) and Local Binary Patterns (LBP), initial defect areas are preliminary localized through image block and unit segmentation, feature extraction and fusion, and dimensionality reduction via Principal Component Analysis (PCA). Based on polar coordinate system and monocular camera ranging theory, the circumferential and axial distance from the reference points to defects are calculated, in order to achieve precise positioning of pipeline defects. The experimental results show that the positioning method has a circumferential position error ≤ 5.00% and an axial position error ≤ 0.10m, with high positioning accuracy. It can provide quantifiable analysis conditions for early warning, positioning, and maintenance of urban drainage pipeline defects.