Experiment results for the paper "Uncertainty-Aware Ship Location Estimation using Multiple Cameras in Coastal Areas" to appear in MDM'2024
收藏NIAID Data Ecosystem2026-05-01 收录
下载链接:
https://zenodo.org/record/10932210
下载链接
链接失效反馈官方服务:
资源简介:
After decompression, there are 16 folders which corresponding to the 16 multi-camera settings in the paper.
Under each folder, there are two files: trajs.csv and trajsGuess.csv.
1. trajs.csv contains the trajectories of ships that are located inside the monitored area of the mult-camera setting.
The first four columns are MMSI (ship identity), timestamp, lon, and lat.
The following columns are the corresponding pixel of the coordinate (lon, lat) in each camera, where (-1,-1) means (lon, lat) is outside the monitored area by a camera.
A pixel is a pair of integers.
xPos1 and yPos1 are for the 1st camera, and xPos2 and yPos2 are for the 2nd camera, and so on so forth.
2. trajsGuess.csv contains the estimated ship locations by using the proposed approach in the paper.
There are 6 columns.
The 1st column is timestamp.
The 2nd column is used to distinguish between the different pixel polygon intersections.
The 3rd column and the 4th column can be either a pixel coordinate or a spatial point coordinate in lon/lat.
The 5th column is either the cameraID of a pixel, or the order of a boundary point for a spatial polygon. The cameraID starts from 1.
The 6th column is the type of the record, which can be
pixel,
or intersection1 (a polygon),
or center1 (center of intersection1),
or intersection2 (a polygon),
or center2 (center of intersection2).
Note that intersection2 and center2 appear rarely in the 6th column.
创建时间:
2024-04-05



