Regular Papers

International Journal of Control, Automation, and Systems 2023; 21(11): 3734-3745

https://doi.org/10.1007/s12555-022-0764-6

© The International Journal of Control, Automation, and Systems

Enhanced Airborne Optical Sectioning Design via HSV Color Space for Detecting Human Object Under Obscured Aerial Image Environment

KangSoo Ryu, Byungjin Lee, Dong-Gyun Kim, and Sangkyung Sung*

Konkuk University

Abstract

This paper aims to develop an algorithm to identify the obscured target from the aerial image encountered in a general complex flight environment. For this, an airborne optical sectioning (AOS) and synthesizing technique is newly adapted for the purpose of preliminary data processing. Specifically, the proposed algorithm consists of several image processing stages. In a foliage-rich environment, the majority of an image may be occupied by trees, where it is difficult to detect a person hidden under foliage. To resolve this, the preprocessing stage in the HSV color space removes the tree parts while preserving the person, and the processed images are combined to construct the AOS, which reveals the person hidden under the trees. Then the exposed person is identified using the machine learning technique that accurately recognizes human shapes. Consequently, the proposed optical sectioning and synthetic process suggests to expand the application field of typical vision-based mission system under complex airborne environment. The performance of the proposed algorithm is demonstrated through the airborne data from a high fidelity process-in-the-loop simulator, where sensor and vision measurement are constructed and provided to a level of practical flight environment.

Keywords Aerial image, airborne optical sectioning, high-fidelity simulator, machine learning, vision-based mission system.

Article

Regular Papers

International Journal of Control, Automation, and Systems 2023; 21(11): 3734-3745

Published online November 1, 2023 https://doi.org/10.1007/s12555-022-0764-6

Copyright © The International Journal of Control, Automation, and Systems.

Enhanced Airborne Optical Sectioning Design via HSV Color Space for Detecting Human Object Under Obscured Aerial Image Environment

KangSoo Ryu, Byungjin Lee, Dong-Gyun Kim, and Sangkyung Sung*

Konkuk University

Abstract

This paper aims to develop an algorithm to identify the obscured target from the aerial image encountered in a general complex flight environment. For this, an airborne optical sectioning (AOS) and synthesizing technique is newly adapted for the purpose of preliminary data processing. Specifically, the proposed algorithm consists of several image processing stages. In a foliage-rich environment, the majority of an image may be occupied by trees, where it is difficult to detect a person hidden under foliage. To resolve this, the preprocessing stage in the HSV color space removes the tree parts while preserving the person, and the processed images are combined to construct the AOS, which reveals the person hidden under the trees. Then the exposed person is identified using the machine learning technique that accurately recognizes human shapes. Consequently, the proposed optical sectioning and synthetic process suggests to expand the application field of typical vision-based mission system under complex airborne environment. The performance of the proposed algorithm is demonstrated through the airborne data from a high fidelity process-in-the-loop simulator, where sensor and vision measurement are constructed and provided to a level of practical flight environment.

Keywords: Aerial image, airborne optical sectioning, high-fidelity simulator, machine learning, vision-based mission system.

IJCAS
July 2024

Vol. 22, No. 7, pp. 2055~2340

Stats or Metrics

Share this article on

  • line

Related articles in IJCAS

IJCAS

eISSN 2005-4092
pISSN 1598-6446