| Title |
Deep Learning based Disparity Map Estimation using Stereo Vision for UAV |
| Authors |
이예지(Yegi Lee) ; 윤경로(Kyoungro Yoon) |
| DOI |
https://doi.org/10.5370/KIEE.2020.69.5.723 |
| Keywords |
Stereo vision; UAV; Semi-supervised learning; Disparity map; Obstacle detection |
| Abstract |
Recently, UAVs(Unmanned Aerial Vehicles) are used in various industries such as military, transportation, agriculture and reconnaissance. However, it is very likely to cause an accident such as a collision or fall, due to external environmental factors, and research is needed to increase safety. To prevent such risks, UAVs are often equipped with sensors such as laser scanners or cameras. But laser scanners are very heavy and consume high power. Stereo cameras are much lighter and use less power than laser scanners, making them ideal for use in small UAV environments. Therefore, in this paper, we introduce a method for estimating the disparity map using a stereo camera and deep learning without using a LiDAR(Lighting Detection And Ranging). The proposed method constructs semi-supervision based neural network to estimate disparity maps. This algorithm can estimate more precise disparity maps than existing matching algorithms. |