Image-objects Manipulation Engine (IoME): semi-automated feature extraction from Very High Resolution (VHR) remotelysensed imagery

Features identification and extraction from remotely sensed (RS) image is an ongoing research endeavor and has wider applications. Traditionally it has been based on pixel-based image analysis which has proved to be inefficient and ineffective especially for very high resolution (VHR) data. More recently object-based image analysis (OBIA) has gained a wider recognition because of its potential for accurately extracting objects from RS data corresponding to real-world features. However OBIA faces various challenges including: the complexity of the algorithms, parameterization and rules development (mostly manual trial and error based), accuracy assessment, and the dependency on a skilled human analyst (error prone and inconsistent). The proposed research, Image-objects Manipulation Engine (IoME), fills the immediate need in the area of automating the feature identification and extraction from VHR image data. The project is envisaged to contribute towards PCI Geomatics’s initiative for developing a tool kit for accurate object extraction from images.

Intern: 
Masroor Hussain
Superviseur universitaire: 
Dr. DongMei Chen
Project Year: 
2014
Province: 
Ontario
Université: 
Partenaire: 
Programme: