___________________________
Corr
espondence author, *mingdelie@yahoo.com; phone 86 027 87540164
A novel method of ship target detection based on
sea-sky-line region extraction
Yongfang Zhang
a
, XiaoSun
a
, Bin Li
a
, and Delie Ming
a*
a
State Key Laboratory for Multi-spectral Information Processing Technology, Institute for Pattern
Recognition and Artificial Intelligence, Huazhong University of Science and Technology, Wuhan
430074, China, Wuhan 430074, China
ABSTRACT
The ship target in the long-range imaging is lack of the features such as size, texture ,shape and so on, so the ship target
detection in the context of complex sky and sea has much difficulty. However, the character that the ship target is always
located in the sea-sky area will help narrow down the search and suppress the unnecessary noise
interference outside the
area. The paper proposes a novel method of ship target detection based on sea-sky-line Detection. The method first use a
robust and efficient method to extract the sea-sky-line region based on segmentation result, Then, we design a
multi-scale filter and adopt an adaptive threshold method of image segmentation to extract the targets from the filtered
image. Our filter is adaptive to the variable size of targets and has good ability to overcome the noise. The experimental
results show that this method can detect IR target such as warship in a complex sea background.
Keywords: target detection, segmentation, sea-sky-line region, multi-scale filter
INTRODUCTION
Automatic target recognition and tracking is a major component of precision-guided systems, and it extracts objects of
interest from the image and uses characteristics of the target for target tracking. ATR is a key component of modern
military defense and attack system. Target detection is critical to tracking, and the subsequent identification and tracking
process can be carried out smoothly only detecting the target accurately[1-2]. So we can say that the target detection
system directly determines the accuracy of the missile strike and demand range.
However, Many factors casts a negative light on the infrared target detection system in the context of complex sky and
sea such as atmospheric attenuation during in the long-range transmission, inhomogeneous thermal radiation of the sky
and the sea , and the noise of infrared sensor and so on. These interference effects cause image signal is in low SNR , the
blurred edge , and lack of texture information ,etc. and what's worse, some parts or all of the target may be hidden in the
background. These make infrared target detection very difficult.
The development of high-performance infrared ship target detection system in sky and sea background is an important
means to improve the defense capabilities, therefore, the study of infrared ship target detection has important theoretical
and practical significance.
The paper is organized as follows: In section 2, commonly used infrared target detection methods are introduced. In
section 3, a novel infrared target detection method adopted in this paper to extract the ship target is described. We
introduce our approach in three phases : graph-based segmentation, extracting sea-sky-line regions, multi-scale filters.
Next, in section 4, the complete experiment of the proposed algorithm are carried out, and the experiment results are also
given out. Finally, in section 5, we will conclude our works.
RELATED WORK
We review prior work along two major themes: sea-sky-line region extraction, and infrared target detection.
1.1 Sea-sky-line Region Extraction
The detection of the sky-sea-line plays an important role in the ship target detection in the infrared image. Sea-sky-line
was determined as position of region which has largest row average gradient in [3], Zhang Feng, etc, used Row-Mapping
Histogram algorithm to directly detect the sea-sky-line area, these two methods is only suitable for very simple sea and
MIPPR 2013: Automatic Target Recognition and Navigation, edited by Tianxu Zhang, Nong Sang, Proc. of SPIE
Vol. 8918, 89181B · © 2013 SPIE · CCC code: 0277-786X/13/$18 · doi: 10.1117/12.2031524
Proc. of SPIE Vol. 8918 89181B-1