Benchmark dataset for small and narrow rectangular object detection from Google Earth imagery

Citation Author(s):
shenlu
jiang
Sungkyunkwan University
zhonghua
hong
Shanghai Ocean University
Submitted by:
zhonghua hong
Last updated:
Tue, 05/17/2022 - 22:17
DOI:
10.21227/j7nx-2495
Data Format:
Research Article Link:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

The benchmark dataset  are consisted of 2,413 three-channel RGB images obtained from Google Earth satellite images and AID dataset.

Instructions: 

To extract the dataser, please unzip the main file 'small-narrow-rectangular-object-7.zip'. There are 3 folders inside it: Annotations contains the Voc format of the labels. JPEGImages contains the original images. Labels contains the intial labeled file. All images are in JPEG format.