Windows and Doors Extraction from Point Cloud Data Combining Semantic Features and Material Characteristics

Point cloud data have become the primary spatial source for 3D reconstruction of building engineering, where reconstructed information models can improve construction efficiency. In such applications, detecting windows and doors is essential. Previous research mainly used red-green-blue (RGB) or semantic features detection, combination these two was not considered. Therefore, this proposed a practical approach to using point with material characteristics. The are first segmented Gradient Filtering Random Sample Consensus (RANSAC) obtain indoor without intrusions protrusions. As input, projected horizontal planes as 2D data. then transformed images, representing area feature extraction. On boundary each potential opening extracted an improved Bounding Box algorithm, extraction result back Based on data, reflectivity applied differentiate from openings, number points check condition doors. abovementioned tested one campus building, including big rooms corridor. experimental results showed that accurate detection successfully reached. completeness 100%, correctness 90.32%. total time 22.8 s processing 2 million reading 10.319 showing 4.938 s.