Jilin He

The detection of crack defects on the walls of road tunnels is a crucial step in the process of ensuring travel safety and performing routine tunnel maintenance. The automatic and accurate detection of cracks on the surface of road tunnels is the key to improving the maintenance efficiency of road tunnels. Machine vision technology combined with a deep neural network model is an effective means to realize the localization and identification of crack defects on the surface of road tunnels. We propose a complete set of automatic inspection methods for identifying cracks on the walls of road tunnels as a solution to the problem of difficulty in identifying cracks during manual maintenance. First, a set of equipment applied to the real-time acquisition of high-definition images of walls in road tunnels is designed. Images of walls in road tunnels are acquired based on the designed equipment, where images containing crack defects are manually identified and selected. Subsequently, the training and validation sets used to construct the crack inspection model are obtained based on the acquired images, whereas the regions containing cracks and the pixels of the cracks are finely labeled. After that, a crack area sensing module is designed based on the proposed you only look once version 7 model combined with coordinate attention mechanism (CA-YOLO V7) network to locate the crack regions in the road tunnel surface images. Only subimages containing cracks are acquired and sent to the multiscale semantic segmentation module for extraction of the pixels to which the cracks belong based on the DeepLab V3+ network. The precision and recall of the crack region localization on the surface of a road tunnel based on our proposed method are 82.4% and 93.8%, respectively. Moreover, the mean intersection over union (MIoU) and pixel accuracy (PA) values for achieving pixel-level detection accuracy are 76.84% and 78.29%, respectively. The experimental results on the dataset show that our proposed two-stage detection method outperforms other state-of-the-art models in crack region localization and detection. Based on our proposed method, the images captured on the surface of a road tunnel can complete crack detection at a speed of ten frames/second, and the detection accuracy can reach 0.25 mm, which meets the requirements for maintenance of an actual project. The designed CA-YOLO V7 network enables precise localization of the area to which a crack belongs in images acquired under different environmental and lighting conditions in road tunnels. The improved DeepLab V3+ network based on lightweighting is able to extract crack morphology in a given region more quickly while maintaining segmentation accuracy. The established model combines defect localization and segmentation models for the first time, realizing pixel-level defect localization and extraction on the surface of road tunnels in complex environments, and is capable of determining the actual size of cracks based on the physical coordinate system after camera calibration. The trained model has high accuracy and can be extended and applied to embedded computing devices for the assessment and repair of damaged areas in different types of road tunnels.

Investigating the path planning of multiple unmanned aerial vehicles (UAVs) covering multiple regions, this work proposes an effective heuristic method of region coverage path planning to reduce the complexity of the problem. The proposed method decomposes the solution process into two stages. First of all, the two most important parameters affecting the performance of UAV missions were considered, namely, the flying speed and the scan width. According to these two parameters of UAVs, a new multi-regional allocation scheme based on the minimum consumption ratio was proposed. With this allocation scheme, the coverage task allocation and path pre-planning of UAVs were obtained. Then, the UAVs’ trajectory routes were optimized based on the dynamic planning algorithm to reduce the time consumption of UAVs on the transfer path between regions. The method was evaluated with numerical experiments. The results showed that the proposed method can effectively solve the path planning problem of multiple UAVs covering multiple regions. Compared with an advanced algorithm, the time consumption for homogeneous and heterogenous UAV performance was reduced by 5.1% and 3%, respectively.

SLAM (Simultaneous Localization and Mapping) based on lidar is an important method for UGV (Unmanned Ground Vehicle) localization in real time under GNSS (Global Navigation Satellite System)-denied situations. However, dynamic objects in real-world scenarios affect odometry in SLAM and reduce localization accuracy. We propose a novel lidar SLAM algorithm based on LOAM (Lidar Odometry and Mapping), which is popular in this field. First, we applied elevation maps to label the ground point cloud. Then we extracted convex hulls in point clouds based on scanlines as materials for dynamic object clustering. We replaced these dynamic objects with background point cloud to avoid accuracy reduction. Finally, we extracted feature points from ground points and non-ground points, respectively, and matched these feature points frame-to-frame to estimate ground robot motion. We evaluated the proposed algorithm in dynamic industrial park roads, and it kept UGV maximum relative position error less than 3% and average relative position error less than 2%.

This paper addresses the design and application controller for a small-size unmanned aerial vehicle (UAV). In this work, the main objective is to study the modeling and attitude controller design for a small size helicopter. Based on a non-simplified helicopter model, a new robust attitude control law, which is combined with a nonlinear control method and a model-free method, is proposed in this paper. Both wind gust and ground effect phenomena conditions are involved in this experiment and the result on a real helicopter platform demonstrates the effectiveness of the proposed control algorithm and robustness of its resultant controller.