A GUI based application for Low Intensity Object Classification & Count using SVM Approach

There is a requirement of processing low intensity images from EO (Electro Optical), IR (Infra-Red) and ISAR (Inverse Synthetic Aperture Radar) sensors on airborne platforms to detect and classify targets(ships, vessels, other objects) against a library of images in real time with a degree of confid...

Full description

Saved in:
Bibliographic Details
Main Authors: Gupta, Vishal, Marriwala, Nikhil, Gupta, Monish
Format: Conference Proceeding
Language:English
Subjects:
CNN etc
Computer vision
Libraries
Machine learning
Object detection
Optical imaging
ships detections
Signal processing
Signal processing algorithms
Support vector machines
SVM
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:There is a requirement of processing low intensity images from EO (Electro Optical), IR (Infra-Red) and ISAR (Inverse Synthetic Aperture Radar) sensors on airborne platforms to detect and classify targets(ships, vessels, other objects) against a library of images in real time with a degree of confidence. Pre-processing of the test input images improves the accuracy of detection and classification. The proposed method verifies and validates the trained software against the pre-processed images. The objective of the proposed approach is to train the machine learning technique in order to estimate the efficiency and accuracy of the classified and detected output from the Deep leaning models. In our work, we have compared the results in terms of accuracy and time with the previous researchers work and we have summarized about our method gives better accuracy.
ISSN:2643-8615
DOI:10.1109/ISPCC53510.2021.9609470