Assisted approaches for forest and non-forest ground truth annotation in satellite remote sensing images / Imran Md Jelas, Mohd Asyraf Zulkifley and Mardina Abdullah

Md Jelas, Imran and Zulkifley, Mohd Asyraf and Abdullah, Mardina (2024) Assisted approaches for forest and non-forest ground truth annotation in satellite remote sensing images / Imran Md Jelas, Mohd Asyraf Zulkifley and Mardina Abdullah. In: UNSPECIFIED.

Abstract

This paper presents a methodology for enhancing ground truth annotations in satellite remote sensing images using Python and OpenCV's SimpleBlobDetector. Focusing on a dataset from Chini Lake, Malaysia, the study adjusts threshold values and employs additional algorithms to reduce noise. The findings demonstrate significant accuracy improvements, with an average of 97.60% using the DeepLabV3+ algorithm. This paper highlights the importance of robust ground truth annotations for ecological monitoring and suggests future research directions.

Metadata

Item Type: Conference or Workshop Item (Paper)
Creators:
Creators
Email / ID Num.
Md Jelas, Imran
imran499@uitm.edu.my
Zulkifley, Mohd Asyraf
UNSPECIFIED
Abdullah, Mardina
UNSPECIFIED
Contributors:
Contribution
Name
Email / ID Num.
Chief Editor
Abdul Rahman, Zarinatun Ilyani
UNSPECIFIED
Editor
Mohd Nasir, Nur Fatima Wahida
UNSPECIFIED
Editor
Kamarudin, Syaza
UNSPECIFIED
Designer
Ramlie, Mohd Khairulnizam
UNSPECIFIED
Subjects: T Technology > T Technology (General) > Technological change > Technological innovations
Divisions: Universiti Teknologi MARA, Perak > Seri Iskandar Campus > Faculty of Architecture, Planning and Surveying
Journal or Publication Title: 13th International Innovation, Invention & Design Competition (INDES 2024)
Page Range: pp. 114-117
Keywords: ground truth, forest, non-forest, forest monitoring, annotation
Date: 2024
URI: https://ir.uitm.edu.my/id/eprint/105415
Edit Item
Edit Item

Download

[thumbnail of 105415.pdf] Text
105415.pdf

Download (2MB)

ID Number

105415

Indexing

Statistic

Statistic details