Skip to content

supervisely-ecosystem/export-to-dota

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Export to DOTA

OverviewHow To Use

GitHub release (latest SemVer) views runs

Overview

Converts project in Supervisely format to DOTA and prepares downloadable archive.

DOTA is one of the formats supported by mmrotate GitHub Org's stars

Preparation

Project object classes shapes must be: Polygon, Bitmap, Line, Rectangle or Any Shape with any of the mentioned shapes, all other shapes will be ignored, and will not be presenting in resulted project. You can convert object classes shapes using convert-class-shape application.

Output archive structure

archive.tar
├── dataset_name_1
│   ├── images
│   └── labelTxt
└── dataset_name_2
    ├── images
    └── labelTxt

Output annotation format

IMAGE_001.txt

First 2 lines of the annotaion is image metadata.

Other lines in the annotation file are labels: x1, y1, x2, y2, x3, y3, x4, y4, label_name complexity

imagesource:imagesource
gsd:None
388.0 1190.0 473.0 1144.0 488.0 1171.0 403.0 1217.0 truck 0
575.0 1024.0 591.0 995.0 696.0 1053.0 679.0 1082.0 truck 0
709.0 1047.0 753.0 1045.0 754.0 1063.0 710.0 1065.0 car 0
714.0 1005.0 714.0 980.0 764.0 980.0 764.0 1005.0 car 0
710.0 923.0 752.0 923.0 752.0 942.0 710.0 942.0 car 0
1125.0 933.0 1171.0 919.0 1176.0 936.0 1130.0 950.0 car 0
380.0 1475.0 519.0 1351.0 541.0 1376.0 402.0 1499.0 truck 0

How to use

App can be launched from ecosystem, images project or images dataset

Run from Ecosystem

Step 1. Run the app from Ecosystem

Step 2. Select input project or dataset and press the Run button

Run from Images Project or Dataset

Step 1. Run the application from the context menu of the Images Project or Dataset

Step 2. Press the Run button

Result