Spatio-Temporal Saliency Fusion Based Small Infrared Moving Target Detection Under Sea-Sky Background

Li Shaoyi,Wang Xiaotian,Zhang Kai,Niu Saisai,Zou Yijun 제어로봇시스템학회 2019 제어로봇시스템학회 국제학술대회 논문집 Vol.2019 No.10

Small infrared moving target detection has an important role in the sea-based infrared search and tracking, maritime area surveillance and other applications. This method aims to detect the small infrared moving targets with the sea-sky background. The present study proposes a detection algorithm for small infrared targets based on the spatio-temporal saliency fusion. The contourlet analysis and edge extraction are carried out in the concurrent design. In order to effectively suppress the background and improve the target signal-clutter ratio, the spectral residual method is combined with the abovementioned methods to reconstruct the target fusion saliency image. Then the target motion region is estimated based on the optical flow method for the fusion saliency image and it is matched with the target area of interest to achieve the moving target detection. Moreover, the pipeline filtering is introduced to achieve the target confirmation by multi-frame judgment, reduce false alarm rate and complete the moving target detection for the infrared image sequence. Experimental results show that the proposed algorithm can achieve continuous target detection and have a higher detection precision via real long wave infrared image sequences.

예측 후보 영역에서의 지역적 대비 차 계산 방법을 활용한 실시간 소형 표적 검출

반종희(Ban Jong-Hee),왕지현(Wang Ji-Hyeun),이동화(Lee Donghwa),유준혁(Yoo Joon-Hyuk),유성은(Yoo Seong-eun) 한국산업정보학회 2017 한국산업정보학회논문지 Vol.22 No.2

본 논문에서는 낮은 SNR을 가지는 적외선 영상에서 강인한 소형 표적 검출을 위해 모폴로지차 연산을 수행하여 표적 후보 영역을 찾고 화소 라벨링을 통해 후보 영역의 위치를 찾는다. 기존의 모폴로지 연산 기반의 표적 검출 방법들은 적외선 영상에 존재하는 클러터에 취약하다는 단점으로 인해 검출정확도가 낮다. 이러한 문제를 해결하기 위해 본 논문에서는 후보 영역에서 표적과 배경 잡음을 분류하기 위해 Moravec 알고리즘과 LCM(Local Contrast Measure) 알고리즘을 결합함으로써 표적 향상과 배경잡음 억제를 동시에 달성한다. 또한, 제안하는 알고리즘은 기존에 실시간 표적 검출을 위해 개발되었던 모폴로지 연산과 가우시안 거리 함수를 이용한 표적 검출 방법의 단일 객체에 제한적인 검출 문제를 해결하여 복수 객체를 효율적으로 검출할 수 있다. In This Paper, we find the Target Candidate Region and the Location of the Candidate Region by Performing the Morphological Difference Calculation and Pixel Labeling for Robust Small Target Detection in Infrared Image with low SNR. Conventional Target Detection Methods based on Morphology Algorithms are low in Detection Accuracy due to their Vulnerability to Clutter in Infrared Images. To Address the Problem, Target Signal Enhancement and Background Clutter Suppression are Achieved Simultaneously by Combining Moravec Algorithm and LCM (Local Contrast Measure) Algorithm to Classify the Target and Noise in the Candidate Region. In Addition, the Proposed Algorithm can Efficiently Detect Multiple Targets by Solving the Problem of Limited Detection of a Single Target in the Target Detection method using the Morphology Operation and the Gaussian Distance Function Which were Developed for Real time Target Detection.

DoG 필터와 시각 주의 모델을 이용한 적외선 소형 표적 검출 방법

장경현(Kyung-Hyun Jang),박기태(Ki-Tae Park),문영식(Young-Shik Moon) 한국정보기술학회 2013 한국정보기술학회논문지 Vol.11 No.6

In this paper, in order to efficiently detect small targets in an infrared image under various background environments, a small target detection method using DoG filter and visual attention model is proposed. To this end, in the first step, a local contrast enhancement is performed to improve intensity of dim small targets. In the second step, candidate regions of small target are extracted from an input infrared image by applying a DoG filter(band pass filter). In the third step, the spectral residual technique which is one of the visual attention models is used to calculate a saliency map. The extracted candidate regions are weighed by the saliency map to suppress background clutters. Finally, an adaptive thresholding technique is applied to detect small targets. The experimental results show that proposed method achieves better performance than the existing methods in various background environments.

복잡한 FLIR 영상에서의 소형 표적 탐지 기법

이승익(Seung-Ik Lee),김주영(Ju-Young Kim),김기홍(Ki-Hong Kim),구본호(Bon-Ho Koo) 한국멀티미디어학회 2007 멀티미디어학회논문지 Vol.10 No.4

본 논문에서는 복잡한 배경을 가지는 전방 관측 열상(FUR; forward looking infrared) 영상에서의 소형 표적 탐지 기법을 제안하였다. 제안한 기법에서는 먼저 이전 프레임과 현재 프레임의 차를 구하여 표적의 움직임 정보를 획득할 뿐만 아니라 시간적으로 발생하는 배경 잡음을 제거한다. 이때 먼 거리에서 다가오는 표적이나 속도가 느린 표적의 경우 차 영상 내에서의 표적의 움직임 정보는 매우 작은 명암도 값을 가진다. 이런 작은 명암도 값을 두드러지게 하여 표적 탐지를 용이하게 하기위하여 프레임 차 영상에 국부 감마 교정을 행한다. 이렇게 표적이 개선된 영상에서 국부적인 통계적인 특성을 이용하여 탐지 지표를 계산한 후 가장 낮은 탐지 지표값을 탐지하고자하는 표적으로 선정한다. 실험을 통하여 제안한 기법이 표적의 탐지 성능이 기존의 탐지기법보다 우수하였음을 확인하였다. In this paper, we propose a small target detection algorithm for FUR image with complex background. First, we compute the motion information of target from the difference between the current frame and the created background image. However, the slow speed of target cause that it has the very low gray level value in the difference image. To improve the gray level value, we perform the local gamma correction for difference image. So, the detection index is computed by using statistical characteristics in the improved image and then we chose the lowest detection index a true target. Experimental results show that the proposed method has significantly the good detection performance.

데이터 기반 영역 제안 및 심층 학습 분류 네트워크를 이용한 소형 적외선 드론 검출

류준환(Junhwan Ryu),김성호(Sungho Kim) 제어로봇시스템학회 2018 제어·로봇·시스템학회 논문지 Vol.24 No.12

In this paper, we propose a data-driven and deep-learning based classification scheme for small infrared target detection. Previous studies have shown feasible performance using conventional computer vision techniques, such as spatial and temporal filters. However, those handcrafted approaches are not optimized due to the nature of the application fields. Recently, deep-learning has shown excellent performance for many computer vision problems. The proposed data-driven proposal and convolutional neural network (DDP-CNN) approach can generate possible target locations through the DDP, and final targets are recognized through the CNN for classification. According to the experimental results using drone databases, the DDP-CNN shows an 0.85 average precision (AP) of target detection.

해상 소형표적 탐색 모형 및 작전운영 방안

이호철(Hocheol Lee),이문걸(Moongul Lee) 육군사관학교 화랑대연구소 2021 한국군사학논집 Vol.77 No.3

Currently, Korea Navy and Coast Guard are conducting search and patrol operations in Korean territorial seas based on their experiences and intuitions from now on. However, these operations have failed frequently due to the characteristics of small targets(like small ship or boat etc), and have caused many arguments and restrictions for search and detection operations. Usually, search operations of small targets have many constraints. The search operations work only if weapon systems such as radar and electro-optical/infra-red equipment reach close to small targets. In this study, we propose an effective search model based on MDP(Markov Decision Process) for limited search spaces and small targets. In addition, we consider realistic maritime environmental characteristics such as fishing nets, sea routes of moving large commercial ship, submerged rock, lighthouse and so on in this model. The various experimental results show that the detection probability of proposed model is significantly better than that of a random search model. We are looking forward to applying this scientific search model for the Korea Navy maritime operation planners.

Advancements in Electro-Optical and Infrared Maritime Detection: A Survey

Changyeol Jeong,Jeonghoon Park,Dayeong Kim,Seunghan Kim,Jeongseop Seo,Sungho Kim 제어로봇시스템학회 2023 제어로봇시스템학회 국제학술대회 논문집 Vol.2023 No.10

표준편차 필터와 밀도기반 군집화를 이용한 적외선 영상 소형 단일 표적 탐지 알고리즘

이인호,이재홍,박찬국 제어·로봇·시스템학회 2023 제어·로봇·시스템학회 논문지 Vol.29 No.1

Small target detection in an infrared search and track (IRST) system is challenging because IR imaging lacks shape information, and a low signal-to-noise ratio. The existing small IR target detection methods have achieved high detection performance without considering the execution time. Hence, we propose a fast and powerful single-frame IR small target detection algorithm with low computational cost, while maintaining excellent detection performance. The IR intensity difference value based on the standard deviation is used to speed up the small target detection and improve detection accuracy. Then, density-based clustering helps to detect the shape of an object and can easily identify the centroid point. By integrating these two approaches, the actual target is selected with the smallest sum of intensities within a bounding box of a specific size. The proposed method is a novel small target detection algorithm that differs from the existing ones. We built over 300 datasets with various scenes and compared other algorithms. The experimental results demonstrated that the proposed algorithm is suitable for real-time applications and effective even when the target size is as small as 2 × 2 pixels.

스케일 불변적인 연산량 감소를 위한 경량 실시간 소형 적외선 표적 검출 알고리즘

반종희,유준혁,Ban, Jong-Hee,Yoo, Joonhyuk 대한임베디드공학회 2017 대한임베디드공학회논문지 Vol.12 No.4

Detecting small infrared targets from the low-SCR images at a long distance is very hard. The previous Local Contrast Method (LCM) algorithm based on the human visual system shows a superior performance of detecting small targets by a background suppression technique through local contrast measure. However, its slow processing speed due to the heavy multi-scale processing overhead is not suitable to a variety of real-time applications. This paper presents a lightweight real-time small target detection algorithm, called by the Improved Selective Local Contrast Method (ISLCM), to reduce the scale-invariant computational overhead. The proposed ISLCM applies the improved local contrast measure to the predicted selective region so that it may have a comparable detection performance as the previous LCM while guaranteeing low scale-invariant computational load by exploiting both adaptive scale estimation and small target feature feasibility. Experimental results show that the proposed algorithm can reduce its computational overhead considerably while maintaining its detection performance compared with the previous LCM.

Real-time small target detection method Using multiple filters and IPP Libraries in Infrared Images

Chul Joong Kim(김철중),Jae Hyup Kim(김재협),Kyung Hyun Jang(장경현) 한국컴퓨터정보학회 2016 韓國컴퓨터情報學會論文誌 Vol.21 No.8

In this paper, we propose a fast small target detection method using multiple filters, and describe system implementation using IPP libraries. To detect small targets in Infra-Red images, it is mandatory that you should apply a filter to eliminate a background and identify the target information. Moreover, by using a suitable algorithm for the environments and characteristics of the target, the filter must remove the background information while maintaining the target information as possible. For this reason, in the proposed method we have detected small targets by applying multi area(spatial) filters in a low luminous environment. In order to apply the multi spatial filters, the computation time can be increased exponentially in case of the sequential operation. To build this algorithm in real-time systems, we have applied IPP library to secure a software optimization and reduce the computation time. As a result of applying real environments, we have confirmed a detection rate more than 90%, also the computation time of the proposed algorithm have been improved about 90% than a typical sequential computation time.