Increasingly very-high resolution (VHR) images become available from camera systems attached to an Unmanned Aerial Vehicle (UAV). For optimal use of those images often traditional pixel based classification or parameter retrieval techniques do not use the full spatial dimension of the provided imagery.
Therefore new image processing methods need to be developed and evaluated to use the full spatial resolution capacity of these VHR images.
Due to the increasingly availability of very-high resolution images, the domain of Object-based Image Analysis (OBIA) and machine vision is growing rapidly. Applications can be found in agro-food analysis, and security (face recognition). In this study, we would like to explore and evaluate the added value of these techniques for the application of identification of plant objects with the final goal to improve retrieval of vegetation characteristics.
However, this study will focus on identification of individual plant objects identifying them from the background and the neighbouring plants. After selection of appropriate OBIA techniques (e.g., Random Forest and others) different VHR datasets are available to evaluate the application on plant objects in potato and banana cropping systems.
- Make an inventory and select relevant Object-based Image Analysis (OBIA) techniques for the analysis of remote sensing based VHR images
- Implement and evaluate a selection of OBIA techniques to identify plant objects in different cropping systems (e.g., potato, banana)
- Evaluate the accuracy of the OBIA techniques for plant object identification.
- Quanlong Feng, Jiantao Liu and Jianhua Gong (2015) UAV Remote Sensing for Urban Vegetation Mapping Using Random Forest and Texture Analysis, Remote Sensing 2015, 7(1), 1074-1094.
- Interest in exploring other domains (e.g., food science, face recognition) where OBIA and machine vision techniques are already used.
Themes: Sensing & measuring, Integrated Land Monitoring