Automatic Extraction of Indoor Structural Information from Point Clouds

Abstract

We propose an innovative method with which to extract building interior structure information automatically, including ceiling, floor, and wall. Our approach outperforms previous methods in the following respects. First, we propose an approach based on principal component analysis (PCA) to find the ground plane, which is regarded as the new Cartesian plane. Second, to reduce the complexity of data processing, the data are projected into two dimensions and transformed into a binary image via the operation of an improved radius outlier removal (ROR) filter. Third, a traditional thinning algorithm is adopted to extract the image skeleton. Then, we propose a method for calculating slope through the nearest neighbor point. Moreover, the line is represented with the slopes to obtain information pertaining to the interior planes. Finally, the outline of the line is restored to a three-dimensional structure. The proposed method is evaluated in multiple scenarios, and the results show that the method is accurate (the maximum error of 0.03 m was in three scenarios) in indoor environments.

Hybrid 3D Reconstruction of Indoor Scenes Integrating Object Recognition

“…It is important to note that the desired property of manifold and watertight model is characterized by the condition that each edge connects only two adjacent faces. This is exemplified by the connections between faces 1 ⃝ and 2 ⃝, 2 ⃝ and 3 ⃝, and 3 ⃝ and 4 ⃝ in Figure 4. Employing the aforementioned energy terms, we can formulate the final optimization function for face selection as follows:…”

“…Based on the extracted geometric primitives, Li et al [7] and Lin et al [18] involve decomposing the point cloud and fitting segmented building blocks together to create rough building models. Cheng et al [1] propose a method that involves projecting 3D point clouds onto 2D and extracting indoor structural information from the resulting 2D images. However, due to the complex and occlusion-rich nature of indoor scenes, this projection-based method is not suitable for representing indoor instance objects.…”

“…It is worth noting that these subsidiary faces are not supported by the points of the origin point cloud; therefore, they can be easily discarded in the following selection. It is evident that among the four faces that share an edge e, the model should be composed of faces 1 ⃝ and 2 ⃝ rather than 1 ⃝ and 3 ⃝, signifying their non-coplanarity. This information serves as a basis for formulating the face constraints utilized in the selection of the constituent manifold and watertight model.…”

“…Indoor 3D models exhibit considerable potential for various applications, such as indoor positioning, architectural design, and augmented reality. In the past few decades, extensive research has been dedicated to the automated 3D reconstruction of indoor models [1][2][3][4]. In comparison to images, 3D point clouds exhibit superior capabilities in representing object geometry, position, and supplementary information.…”

Procedural Point Cloud Modelling in Scan-to-BIM and Scan-vs-BIM Applications: A Review

“…Planar primitive detection [3,31,32,[36][37][38][39][40]45,47,[50][51][52][53][54][55][56][57][58]] Volumetric primitive fitting [59,60] Mesh-based reconstruction [5,11,30,37,57,58,[61][62][63] Next, an overview of the state of the art on outdoor and indoor modelling will be performed. Table 7 categorizes the target environments addressed in recent literature.…”