Objective Compared to traditional single-modality color imaging or 3D imaging, color 3D imaging can simultaneously provide both spectral reflectance and spatial geometric information, significantly enhancing target discrimination and identification performance. Therefore, it has significant potential in applications of robotic vision, autonomous navigation, ecological monitoring, and scientific research. However, existing color 3D imaging methods face several limitations, including short working distances, low spatial resolution, and vulnerability to environmental interference in complex scenarios. To address these challenges, a novel high-resolution color range-gated 3D imaging method is proposed, which synergistically combines laser range-gated 3D imaging with RGB color imaging. This method utilizes gate viewing to effectively suppress atmospheric backscattering while eliminating non-target foreground/background interference, thereby enabling long-range, high-resolution color 3D reconstruction exclusively within the designated target region.

Methods The color range-gated 3D imaging system consists of a range gated imaging module, a color imaging module and a precision timing control unit. The range gated imaging module is responsible for capturing gated images of the target area at frames A and B, while the color imaging module is used to acquire the corresponding RGB image at frame C. Firstly, the ABC-ABC timing scheme (Fig.1(b)) for the timing control unit is designed to alternately acquire A-frame and B-frame gated images synchronized with the C-frame color image (Fig.1(a)). Secondly, the gated images and color images are matched by using the DeDoDe feature matching algorithm. Thirdly, the gated images are reconstructed into depth images based on the range-intensity correlation 3D imaging algorithm (Fig.2). Subsequently, depth-based segmentation is applied to extract the target region from the color images while eliminating foreground and background interference, yielding clean target-specific color images. Finally, the depth information and RGB color formation of color are fused to generate a high-quality color 3D point cloud representation (Fig.3).

Results and Discussions The proposed color range-gated 3D imaging method has been validated through both indoor and outdoor experiments. First, the experimental results of checkerboard calibration target demonstrate that the DeDoDe feature matching algorithm can accurately match gated and color images without a complex calibration process (Fig.5 and Tab.1). Second, indoor experimental results of the Secchi disk targets are shown in Fig.6. Compared to color 3D point cloud from color LiDAR, the point cloud density of color range-gated 3D imaging is improved by more than 57 times (Fig.8 and Tab.2), and the ranging error is ≤2 cm at a distance of 10 m. Compared with LiDAR, the LiRAI system achieves a 7.58-fold improvement in spatial resolution (Tab.2). Third, outdoor field tests of statue demonstrate that the color range-gated 3D imaging system is capable of acquiring high-quality color 3D point clouds in both daytime and nighttime operations (Fig.8 and Fig.9). Finally, the acquisition of color range-gated 3D point clouds for vegetation and a person wearing camouflage demonstrate the capability of color range-gated 3D imaging to effectively filter out complex backgrounds and segment low-contrast targets (Fig.11).

Conclusions A color range-gated 3D imaging method is proposed, which integrates laser range-gated 3D imaging with RGB color imaging to generate high-quality color 3D point clouds. Based on range-intensity correlation algorithm and gate viewing, the method achieves accurate 3D reconstruction while simultaneously filtering out foreground and background regions outside the target area. Exploiting the image-level resolution of dual sensors, it achieves heterogeneous data fusion and generates color 3D point cloud of targets. Experimental results demonstrate that the proposed method can acquire high-resolution, color range-gated 3D point clouds under both daytime and nighttime conditions. This method significantly improves the quality of color point cloud data, thereby facilitating more accurate target detection and recognition.