No abstract

No abstract

No abstract

We present a visual localization framework aided by novel deep attention aware features for autonomous driving that achieves centimeter level localization accuracy. Conventional approaches to the visual localization problem rely on handcrafted features or human-made objects on the road. They are known to be either prone to unstable matching caused by severe appearance or lighting changes, or too scarce to deliver constant and robust localization results in challenging scenarios. In this work, we seek to exploit the deep attention mechanism to search for salient, distinctive and stable features that are good for longterm matching in the scene through a novel end-to-end deep neural network. Furthermore, our learned feature descriptors are demonstrated to be competent to establish robust matches and therefore successfully estimate the optimal camera poses with high precision. We comprehensively validate the effectiveness of our method using a freshly collected dataset with high-quality ground truth trajectories and hardware synchronization between sensors. Results demonstrate that our method achieves a competitive localization accuracy when compared to the LiDAR-based localization solutions under various challenging circumstances, leading to a potential low-cost localization solution for autonomous driving. IntroductionLocalization is a fundamental task in a self-driving car system. To exploit high definition (HD) maps as priors for robust perception and safe motion planning, this requires the localization system to reach centimeter-level accuracy [2].Despite many decades of research, building a long-term, precise and reliable localization system using low-cost sensors, such as automotive and consumergrade GPS/IMU and cameras, is still an open-ended and challenging problem. Compared to the LiDAR, that cameras are passive sensors meaning that they are more susceptible to appearance changes caused by varying lighting conditions or changes in viewpoint. It is known that handcrafted point features (DIRD [26,28],

Abstract. The resolution of geostationary satellite image is not high and the image is covered with clouds. At present, when the extracted feature points are unstable, there are some problems, such as low matching accuracy or even matching failure. In this paper, a landmark matching algorithm is proposed to directly establish the multi-level grids for the image coastline and the coastline template. Through the similarity measure of the multi-level grids, the landmark matching is realized layer by layer. First of all, we've finished cloud detection, establishment of landmark data set, and extraction of image coastline. Then we design and implement the landmark matching algorithm based on multi-level grids. Finally, through analysis from different levels of landmarks and different proportion of cloud cover, the advantages and applicable conditions of this algorithm are given. The experimental results show that: 1) with the increase of cloud cover, the correct rate of landmark matching decreases, but the decrease is small. It shows that the matching algorithm in this paper is stable. Correct matching rate could always be stable at about 75 percent in the fourth level. 2) when the proportion of cloud cover is less than 20 percent, the higher the matching level, the higher the matching accuracy. When the cloud cover is more than 20 percent, the matching accuracy in the fourth level is the highest. This algorithm provides a stable method for the landmark matching of geostationary satellite image.