In road scenes, the collected 3D point cloud data is usually accompanied by a large amount of interference mainly composed of ground point clouds and the property of uneven density distribution, which will bring difficulties to subsequent recognition and prediction. To address these problems, this paper proposes a ground point cloud segmentation algorithm based on local density plane fitting. Firstly, for the uneven density distribution of 3D point clouds, density segmentation is used to obtain several regions with balanced density. Then, candidate sample selection and plane validity detection are carried out for each region. The modified classical DBSCAN clustering algorithm is used to obtain effective fitting planes and perform clustering according to the fitting planes. Finally, different planes are divided according to the clustering results, and abnormal inspection is performed on the obtained results to screen out the most reasonable result. This scheme can effectively improve the scalability of the algorithm, reduce training costs, and improve deployment efficiency and universality. Experimental results show that the algorithm used in this paper has advantages compared with advanced algorithms of the same category, and can greatly reduce ground interference.

To thoroughly explore the microstructural characteristics of the interfacial transition zone (ITZ) in fiber-reinforced concrete, carbon fiber reinforced concrete (CFSC), carbon nanotube reinforced concrete (CNSC), and multi-scale carbon fiber reinforced concrete (MCSC) were selected as the primary research objects. The nanoindentation test was employed to determine the elastic modulus and hardness of the ITZ in CFSC, CNSC, and MCSC, and the thickness of the ITZ was quantitatively defined using a statistical threshold criterion. The deconvolution method was used to conduct a comparative investigation of the phase composition and distribution characteristics of the ITZ and cement matrix. Parameter upscaling analyses were carried out based on the Mori-Tanaka model and Voigt-Reuss-Hill algorithm. The results reveal that the thickness of the ITZ in CFSC is approximately 16.0 ± 2.8 μm (95% CI: 13.7–18.3 μm), while that in CNSC and MCSC is 12.0 ± 3.1 μm (95% CI: 8.2–15.8 μm) and 12.0 ± 2.5 μm (95% CI: 9.9–14.1 μm), respectively. Compared with CFSC, the fractions of high-density calcium silicate hydrate (HD C-S-H) and calcium hydroxide (CH) in the ITZ of CNSC and MCSC are significantly increased (one-way ANOVA: p = 0.00053 for HD C-S-H; p = 0.0014 for LD C-S-H; Tukey HSD, CFSC vs. CNSC/MCSC: p < 0.01), and the fraction of low-density calcium silicate hydrate (LD C-S-H) is notably decreased. However, the pore content in CNSC and MCSC is relatively higher, which is attributed to the water-absorption and subsequent release behavior of CN. The main factor contributing to the difference in mechanical properties between the ITZ and cement matrix lies in the different proportions of each constituent phase. The equivalent elastic moduli calculated by the parameter upscaling method are ranked in descending order as CNSC > MCSC > CFSC. The nanoindentation results suggest that CN at the nanoscale may promote hydration and densification of the ITZ, followed by the multi-scale carbon fiber (MC) at the micro-nanoscale, though this interpretation requires confirmation through complementary chemical characterization techniques. The research conclusions provide a scientific basis for understanding the microstructural characteristics and micro-interface strengthening mechanisms of fiber reinforced concrete.