In this letter, we present a long-term localization method that effectively exploits the structural information of an environment via an image format. The proposed method presents a robust year-round localization performance even when learned in just a single day. The proposed localizer learns a point cloud descriptor, named Scan Context Image (SCI), and performs robot localization on a grid map by formulating the place recognition problem as place classification using a convolutional neural network. Our method is faster than existing methods proposed for place recognition because it avoids a pairwise comparison between a query and scans in a database. In addition, we provide thorough validations using publicly available long-term datasets, the NCLT dataset and the Oxford RobotCar dataset, and show that the Scan Context Image (SCI) localization attains consistent performance over a year and outperforms existing methods.

Semantic segmentation is performed to understand an image at the pixel level; it is widely used in the field of autonomous driving. In recent years, deep neural networks achieve good accuracy performance; however, there exist few models that have a good trade-off between high accuracy and low inference time. In this paper, we propose a fast spatial feature network (FSFNet), an optimized lightweight semantic segmentation model using an accelerator, offering high performance as well as faster inference speed than current methods. FSFNet employs the FSF and MRA modules. The FSF module has three different types of subset modules to extract spatial features efficiently. They are designed in consideration of the size of the spatial domain. The multi-resolution aggregation module combines features that are extracted at different resolutions to reconstruct the segmentation image accurately. Our approach is able to run at over 203 FPS at full resolution ( $1024\times 2048$ ) in a single NVIDIA 1080Ti GPU, and obtains a result of 69.13% mIoU on the Cityscapes test dataset. Compared with existing models in real-time semantic segmentation, our proposed model retains remarkable accuracy while having high FPS that is over 30% faster than the state-of-the-art model. The experimental results proved that our model is an ideal approach for the Cityscapes dataset. The code is publicly available at: https://github.com/computervision8/FSFNet .

This paper proposes a hierarchical roadmap (HRM) and its construction process to efficiently represent navigable areas in an indoor environment. HRM is adopted to solve the path‐planning problems of mobile robots in indoor environments. HRM has a multi‐layered graphical structure that enables it to abstract and cover navigable areas using a smaller number of nodes and edges than a probabilistic roadmap. During the incremental process of constructing HRM, information on navigable areas is abstracted using a sonar gridmap when the mobile robot navigates an unexplored area. The HRM‐based planner efficiently searches for paths to answer queries by reducing the search space size using the multi‐layered graphical structure. The benefits of the proposed HRM are experimentally verified in real indoor environments.

Among underwater perceptual sensors, imaging sonar has been highlighted for its perceptual robustness underwater. The major challenge of imaging sonar, however, arises from the difficulty in defining visual features despite limited resolution and high noise levels. Recent developments in deep learning provide a powerful solution for computer-vision researches using optical images. Unfortunately, deep learning-based approaches are not well established for imaging sonars, mainly due to the scant data in the training phase. Unlike the abundant publically available terrestrial images, obtaining underwater images is often costly, and securing enough underwater images for training is not straightforward. To tackle this issue, this paper presents a solution to this field's lack of data by introducing a novel end-to-end image-synthesizing method in the training image preparation phase. The proposed method present image synthesizing scheme to the images captured by an underwater simulator. Our synthetic images are based on the sonar imaging models and noisy characteristics to represent the real data obtained from the sea. We validate the proposed scheme by training using a simulator and by testing the simulated images with real underwater sonar images obtained from a water tank and the sea.

This paper proposes a method to extend a sensing range of a short-baseline stereo camera (SBSC). The proposed method combines a stereo depth and a monocular depth estimated by a convolutional neural network-based monocular depth estimation (MDE). To combine a stereo depth and a monocular depth, the proposed method estimates a scale factor of a monocular depth using stereo depth–mono depth pairs and then combines the two depths. Another advantage of the proposed method is that the trained MDE model may be utilized for different environments without retraining. The performance of the proposed method is verified qualitatively and quantitatively using the directly collected and open datasets.