In practical applications of line-structured light measurement technology, the transmitted laser may be reflected several times due to the reflective properties of the measured object’s surface, which directly affects the accuracy of three-dimensional (3-D) measurements. To obtain precise laser stripes under such reflective conditions to improve measurement accuracy, this paper proposes a semantic segmentation method called GuideNet. Firstly, a dual-branch encoder is introduced to simultaneously extract high-resolution spatial features and high-level semantic features from the input image. Secondly, a Guidance module is designed to fuse spatial features and semantic features and enhance the model’s ability to focus on relevant regions. Finally, the multi-level feature fusion decoder is performed to obtain the final results. Experiments on RIDataset and 3-D measurement experiments demonstrate that the proposed method improves the anti-reflection ability of line-structured light sensors.