Detecting and recognizing objects are crucial steps in interpreting remote sensing images. At present, deep learning methods are predominantly employed for detecting objects in remote sensing images, necessitating a significant number of floating-point computations. However, low computing power and small storage in computing devices are hard to afford the large model parameter quantity and high computing complexity. To address these constraints, this article presents a lightweight detection model called CSPPartial-YOLO. This model introduces the partial hybrid dilated convolution (PHDC) Block module that combines hybrid dilated convolutions and partial convolutions to increase the receptive field at a low computational cost. By using the PHDC Block within the model design framework of cross-stage partial connection, we construct CSPPartialStage that reduces computational burden without compromising accuracy. Coordinate attention module is also employed in CSPPartialStage to aggregate position information and improve the detection of small objects with complex distributions in remote sensing images. A backbone and neck are developed with CSPPartialStage, and the rotation head of the PPYOLOE-R model adapts to objects of multiple orientations in remote sensing images. Empirical experiments using the dataset for object deTection in aerial images (DOTA) dataset and a large-scale small object detection dAtaset (SODA-A) dataset indicate that our method is faster and resource efficient than the baseline model (PPYOLOE-R), while achieving higher accuracy. Furthermore, comparisons with current state-of-the-art YOLO series detectors show our proposed model's competitiveness in terms of speed and accuracy. Moreover, compared to mainstream lightweight networks, our model exhibits better hardware adaptability, with lower inference latency and higher detection accuracy.

Indoor-scene semantic segmentation is of great significance to indoor navigation, high-precision map creation, route planning, etc. However, incorporating RGB and HHA images for indoor-scene semantic segmentation is a promising yet challenging task, due to the diversity of textures and structures and the disparity of multi-modality in physical significance. In this paper, we propose a Cross-Modality Attention Network (CMANet) that facilitates the extraction of both RGB and HHA features and enhances the cross-modality feature integration. CMANet is constructed under the encoder–decoder architecture. The encoder consists of two parallel branches that successively extract the latent modality features from RGB and HHA images, respectively. Particularly, a novel self-attention mechanism-based Cross-Modality Refine Gate (CMRG) is presented, which bridges the two branches. More importantly, the CMRG achieves cross-modality feature fusion and produces certain refined aggregated features; it serves as the most crucial part of CMANet. The decoder is a multi-stage up-sampled backbone that is composed of different residual blocks at each up-sampling stage. Furthermore, bi-directional multi-step propagation and pyramid supervision are applied to assist the leaning process. To evaluate the effectiveness and efficiency of the proposed method, extensive experiments are conducted on NYUDv2 and SUN RGB-D datasets. Experimental results demonstrate that our method outperforms the existing ones for indoor semantic-segmentation tasks.

Tree skeletons play an important role in tree structure analysis and 3D model reconstruction. However, it is a challenge to extract a skeleton from a tree point cloud with complex branches. In this paper, an automatic and fast tree skeleton extraction method (FTSEM) based on voxel thinning is proposed. In this method, a wood–leaf classification algorithm was introduced to filter leaf points for the reduction of the leaf interference on tree skeleton generation, tree voxel thinning was adopted to extract a raw tree skeleton quickly, and a breakpoint connection algorithm was used to improve the skeleton connectivity and completeness. Experiments were carried out in Haidian Park, Beijing, in which 24 trees were scanned and processed to obtain tree skeletons. The graph search algorithm (GSA) was used to extract tree skeletons based on the same datasets. Compared with the GSA method, the FTSEM method obtained more complete tree skeletons. The time cost of the FTSEM method was evaluated using the runtime and time per million points (TPMP). The runtime of FTSEM was from 1.0 s to 13.0 s, and the runtime of GSA was from 6.4 s to 309.3 s. The average value of TPMP was 1.8 s for FTSEM and 22.3 s for GSA, respectively. The experimental results demonstrate that the proposed method is feasible, robust, and fast with good potential for tree skeleton extraction.

Automatically and accurately reconstructing the overhead wires of railway from airborne laser scanning (ALS) data are an efficient way of railway monitoring to ensure stable and safety transportation services. However, due to the complex structure of the overhead wires, it is challenging to extract these wires using the existing methods. This work proposes a workflow for railway overhead wire reconstruction using deep learning for wire identification collaborating with the RANdom SAmple Consensus (RANSAC) algorithm for wire reconstruction. First, data augmentation and ground points down-sampling are performed to facilitate the issues caused by insufficient and non-uniformity of LiDAR points. Then, a network incorporating with PointNet model is proposed to segment wires, pylons and ground points. The proposed network is composed of a Geometry Feature Extraction (GFE) module and a Neighborhood Information Aggregation (NIA) module. These two modules are introduced to encode and describe the local geometric features. Therefore, the capability of the model to discriminate geometric details is enhanced. Finally, a wire individualization and multi-wire fitting algorithm is proposed to reconstruct the overhead wires. A number of experiments are conducted using ALS point cloud data of railway scenarios. The results show that the accuracy and MIoU for wire identification are 96.89% and 82.56%, respectively, which demonstrates a better performance compared to the existing methods. The overall reconstruction accuracy is 96% over the study area. Furthermore, the presented strategy also demonstrated its applicability to high-voltage powerline scenarios.

Hyperspectral LiDAR (HSL) is a promising active detection technique for underwater positioning and remote sensing, enabling the simultaneous acquisition of three-dimensional topographic and spectral information of underwater targets. This study presents an advanced underwater hyperspectral LiDAR (UDHSL) system with a spectral range of 450–700 nm, adjustable spectral bandwidth of 10–300 nm, and tunable repetition frequency of 50 kHz to 1 MHz. The system achieves high precision with a laser divergence angle of ≤1 mrad, pulse width of 7 ns, laser energy of 7.5 µJ, ranging resolution of 1.13 cm and ranging accuracy of 1.02 m@distance of 27 m. Hyperspectral point clouds spanning 11 bands (450–650 nm) are generated during 3D pool experiments. The distance-colored point clouds precisely align with the geometric characteristics of targets, the normalized intensity-colored point clouds across spectral bands exhibit discriminative capabilities for target identification, and the color-composite point clouds approximate the true colors of targets, collectively validating the system’s ability to concurrently acquire spectral and topographic data. These results underscore the potential of this technology for underwater exploration and positioning applications.

The analysis of greenhouse gas emission characteristics on urban roads is of great significance for understanding the sources and sinks of urban greenhouse gases and their changing patterns. Based on tunable diode laser derivative absorption spectroscopy technology, which features high resolution, high sensitivity, and fast response, a vehicle-mounted system capable of simultaneously detecting CO₂ and CH₄ has been developed. The system has a response time of 0.38 s, with detection sensitivities of 5 ppb for CH₄ and 0.2 ppm for CO₂, power consumption of approximately 4.8 W, a weight of less than 3 kg, and dimensions of 255 mm × 275 mm × 85 mm. Using this system, monitoring campaigns were conducted on the same road in Beijing, running north–south, during different time periods in April and June 2023. The results show that there is little correlation between changes in CO₂ and CH₄ concentrations on the road, and these gas concentrations exhibit different influencing factors and spatiotemporal characteristics. The CO₂ concentration on the road is primarily related to the degree of traffic congestion and does not exhibit significant seasonal variations. The average CO₂ concentration measured on the road is much higher than the global average CO₂ concentration during the same period. On the other hand, the CH₄ concentration on the road is not strongly correlated with traffic congestion but is closely related to the leakage of methane from specific emission wells or covers. The CH₄ concentration is higher in the morning, gradually decreases as the sun rises, and then increases again after sunset. The CH₄ concentration measured at night in June is significantly lower than that in April, reflecting some seasonal variation. The CH₄ concentration on the Beijing urban road is slightly higher than the global average CH₄ concentration during the same period. The vehicle-mounted experiments verified the feasibility of using this self-developed system for vehicle-mounted detection of greenhouse gas concentrations on urban roads. The research results can provide data for analyzing the spatial pattern of regional carbon sources and sinks.