Obtaining water body images quickly and reliably is important to guide human production activities and study urban change. This paper presents a fast and accurate method to identify water bodies in complex environments based on UAV high-resolution images. First, an improved U-Net (SU-Net) model is proposed in this paper. By increasing the number of connections in the middle layer of the neural network, more image features can be retained through S-shaped circular connections. Second, aiming at the interference of mixed ground objects and dark ground objects on water detection, the fusion of a deep learning network and visual features is investigated. We analyse the influence of a wavelet transform and grey level cooccurrence matrix (GLCM) on water extraction. Using a confusion matrix to evaluate accuracy, the following conclusions are drawn: (1) Compared with existing methods, the SU-Net method achieves a significant improvement in accuracy, and the overall accuracy (OA) is 96.25%. The kappa coefficient (KC) is 0.952. (2) SU-Net combined with the GLCM has a higher accuracy (OA is 97.4%) and robustness in distinguishing mixed and dark objects. Based on this method, a distinct water boundary in urban areas, which provides data for urban water vector mapping, can be obtained.

With the rapid process of both urban sprawl and urban renewal, large numbers of old buildings have been demolished in China, leading to wide spread construction sites, which could cause severe dust contamination. To alleviate the accompanied dust pollution, green plastic mulch has been widely used by local governments of China. Therefore, timely and accurate mapping of urban green plastic covered regions is of great significance to both urban environmental management and the understanding of urban growth status. However, the complex spatial patterns of the urban landscape make it challenging to accurately identify these areas of green plastic cover. To tackle this issue, we propose a deep semi-supervised learning framework for green plastic cover mapping using very high resolution (VHR) remote sensing imagery. Specifically, a multi-scale deformable convolution neural network (CNN) was exploited to learn representative and discriminative features under complex urban landscapes. Afterwards, a semi-supervised learning strategy was proposed to integrate the limited labeled data and massive unlabeled data for model co-training. Experimental results indicate that the proposed method could accurately identify green plastic-covered regions in Jinan with an overall accuracy (OA) of 91.63%. An ablation study indicated that, compared with supervised learning, the semi-supervised learning strategy in this study could increase the OA by 6.38%. Moreover, the multi-scale deformable CNN outperforms several classic CNN models in the computer vision field. The proposed method is the first attempt to map urban green plastic-covered regions based on deep learning, which could serve as a baseline and useful reference for future research.

The water-level fluctuation zone (WLFZ) of the Three Gorges Reservoir is a serious landslide-prone area. However, current remote sensing methods for landslide mapping and detection in the WLFZ are insufficient because of difficulties in data acquisition and lack of facade information. We proposed a novel shipborne mobile photogrammetry approach for 3D mapping and landslide detection in the WLFZ for the first time, containing a self-designed shipborne hardware platform and a data acquisition and processing workflow. To evaluate the accuracy and usability of the resultant 3D models in the WLFZ, four bundle block adjustment (BBA) control configurations were developed and adopted. In the four configurations, the raw Global Navigation Satellite System (GNSS) data, the raw GNSS data and fixed camera height, the GCPs extracted from aerial photogrammetric products, and the mobile Light Detection and Ranging (LiDAR) point cloud were used. A comprehensive accuracy assessment of the 3D models was conducted, and the comparative results indicated the BBA with GCPs extracted from the aerial photogrammetric products was the most practical configuration (RMSE 2.00 m in plane, RMSE 0.46 m in height), while the BBA with the mobile LiDAR point cloud as a control provided the highest georeferencing accuracy (RMSE 0.59 m in plane, RMSE 0.40 m in height). Subsequently, the landslide detection ability of the proposed approach was compared with multisource remote sensing images through visual interpretation, which showed that the proposed approach provided the highest landslide detection rate and unique advantages in small landslide detection as well as in steep terrains due to the more detailed features of landslides provided by the shipborne 3D models. The approach is an effective and flexible supplement to traditional remote sensing methods.

At present, a common drawback of crowd simulation models is that they are mainly simulated in (abstract) 2D environments, which limits the simulation of crowd behaviors observed in real 3D environments. Therefore, we propose a deep reinforcement learning-based model with human-like perceptron and policy for crowd evacuation in 3D environments (HDRLM3D). In HDRLM3D, we propose a vision-like ray perceptron (VLRP) and combine it with a redesigned global (or local) perceptron (GOLP) to form a human-like perception model. We propose a double-branch feature extraction and decision network (DBFED-Net) as the policy, which can extract features and make behavioral decisions. Moreover, we validate our method’s ability to reproduce typical phenomena and behaviors through experiments in two different scenarios. In scenario I, we reproduce the bottleneck effect of crowds and verify the effectiveness and advantages of HDRLM3D by comparing it with real crowd experiments and classical methods in terms of density maps, fundamental diagrams, and evacuation times. In scenario II, we reproduce agents’ navigation and obstacle avoidance behaviors and demonstrate the advantages of HDRLM3D for crowd simulation in unknown 3D environments by comparing it with other deep reinforcement learning-based models in terms of trajectories and numbers of collisions.

Simultaneous Localization and Mapping (SLAM) has been widely used in emergency response, self-driving and city-scale 3D mapping and navigation. Recent deep-learning based feature point extractors have demonstrated superior performance in dealing with the complex environmental challenges (e.g. extreme lighting) while the traditional extractors are struggling. In this paper, we have successfully improved the robustness and accuracy of a monocular visual SLAM system under various complex scenes by adding a deep learning based visual localization thread as an augmentation to the visual SLAM framework. In this thread, our feature extractor with an efficient lightweight deep neural network is used for absolute pose and scale estimation in real time using the highly accurate georeferenced prior map database at 20cm geometric accuracy created by our in-house and low-cost LiDAR and camera integrated device. The closed-loop error provided by our SLAM system with and without this enhancement is 1.03m and 18.28m respectively. The scale estimation of the monocular visual SLAM is also significantly improved (0.01 versus 0.98). In addition, a novel camera-LiDAR calibration workflow is also provided for large-scale 3D mapping. This paper demonstrates the application and research potential of deep-learning based vision SLAM with image and LiDAR sensors.