True color 3D imaging optimization with missing spectral bands based on hyperspectral LiDAR.

Area of Science:

• Optics and Photonics
• Computer Vision
• Remote Sensing

Background:

• True color 3D imaging is crucial for scene reconstruction and target identification, but traditional methods like fusing 3D point clouds with RGB images face challenges with illumination and registration.
• Hyperspectral LiDAR (HSL) offers a promising active imaging technique for 3D scene reconstruction, avoiding common issues through hardware design.

Purpose of the Study:

• To develop and evaluate a new color reconstruction method for hyperspectral LiDAR (HSL) systems to overcome limitations caused by missing spectral bands.
• To improve the accuracy and reliability of true color 3D imaging when spectral information is incomplete.

Main Methods:

• The study utilized two indoor experiments and evaluated five distinct color reconstruction schemes.
• A novel color reconstruction algorithm was developed and tested to address spectral band deficiencies in HSL data.
• Performance was assessed by comparing the new method against traditional color reconstruction techniques.

Main Results:

• The proposed method demonstrated considerable improvements in both the overall color reconstruction effect and color similarity compared to traditional approaches.
• Color component similarity saw a significant increase, rising from 0.324 to 0.762.
• Imaging results confirmed the method's reliability in enhancing color reconstruction quality even with missing spectral bands.

Conclusions:

• The developed color reconstruction method effectively improves true color 3D imaging with hyperspectral LiDAR (HSL) data, particularly when spectral bands are missing.
• The findings validate the feasibility and repeatability of the new technique, broadening the application potential of HSL measurements in 3D scene analysis.