Study of machine intelligence for SAR image analysis
Image segmentation and object detection are two fundamental but major applications of machine intelligence. With the help of machine learning, image processing work can be done properly in a very short period even if huge amount of images are provided. In this report, the author presents a new ap...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/74573 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Image segmentation and object detection are two fundamental but major applications of machine intelligence. With the help of machine learning, image processing work can be done properly in a very short period even if huge amount of images are provided.
In this report, the author presents a new approach to detect and classify geographical images which are captured by satellites (SAR) by using Extreme Learning Machine (ELM) methodology. Satellite images and extract representative attributes were collected as row sample data. Followed by training the programme with sample data collected and use the trained machine to predict land-cover types of new geographical images.
Compared with conventional segmentation method, which is based on manually work, the performance for machine intelligence auto-classify programme has shown great improvement of time consuming on the test images while maintaining an acceptable accuracy and more consistent performance when dealing with huge amount of images. The introduced machine intelligence based method outperforms the conventional method in both time consuming and labor wasting. |
|---|