《電子技術(shù)應(yīng)用》
您所在的位置:首頁(yè) > 其他 > 設(shè)計(jì)應(yīng)用 > 基于機(jī)載SAR圖像的海面船只實(shí)時(shí)檢測(cè)分類應(yīng)用架構(gòu)
基于機(jī)載SAR圖像的海面船只實(shí)時(shí)檢測(cè)分類應(yīng)用架構(gòu)
2022年電子技術(shù)應(yīng)用第3期
繆煒星1,王小龍2,,李志勇2,,楊 松2,張 欣2,,季昕寧2
1.中航(成都)無(wú)人機(jī)系統(tǒng)股份有限公司,,四川 成都611730;2.中國(guó)科學(xué)院 空天信息創(chuàng)新研究院,,北京100190
摘要: 合成孔徑雷達(dá)(SAR)被廣泛應(yīng)用于軍事偵察,、海洋監(jiān)測(cè)、災(zāi)害應(yīng)急評(píng)估等各類應(yīng)用中,。其中,,SAR海面船只檢測(cè)分類是SAR海洋應(yīng)用的重要一環(huán)。限于SAR特殊的成像機(jī)理,,SAR圖像中的目標(biāo)特征多變,,導(dǎo)致基于SAR圖像的目標(biāo)分類應(yīng)用進(jìn)展緩慢。機(jī)載SAR因其實(shí)時(shí),、便捷的部署和使用方式得到了快速發(fā)展,,基于機(jī)載SAR圖像的海面船只檢測(cè)分類,特別是實(shí)時(shí)檢測(cè)分類應(yīng)用近年來(lái)需求強(qiáng)烈,。為此,,針對(duì)SAR海面船只實(shí)時(shí)檢測(cè)分類應(yīng)用問(wèn)題,探討和構(gòu)建了一種基于機(jī)載SAR圖像的海面船只實(shí)時(shí)檢測(cè)分類系統(tǒng)構(gòu)架,,論述了系統(tǒng)構(gòu)架在工程應(yīng)用中的關(guān)鍵流程,、方法并進(jìn)行了初步試驗(yàn)驗(yàn)證,,結(jié)合當(dāng)前人工智能技術(shù)的發(fā)展趨勢(shì),對(duì)其日后的發(fā)展趨勢(shì)進(jìn)行了展望,。
中圖分類號(hào): TP391.41
文獻(xiàn)標(biāo)識(shí)碼: A
DOI:10.16157/j.issn.0258-7998.211921
中文引用格式: 繆煒星,,王小龍,李志勇,,等. 基于機(jī)載SAR圖像的海面船只實(shí)時(shí)檢測(cè)分類應(yīng)用架構(gòu)[J].電子技術(shù)應(yīng)用,,2022,48(3):113-117.
英文引用格式: Miao Weixing,,Wang Xiaolong,,Li Zhiyong,et al. Real-time ship detection and classification application framework based on airborne SAR image[J]. Application of Electronic Technique,,2022,,48(3):113-117.
Real-time ship detection and classification application framework based on airborne SAR image
Miao Weixing1,Wang Xiaolong2,,Li Zhiyong2,,Yang Song2,Zhang Xin2,,Ji Xinning2
1.Gascoigne Company Limited,,Aviation Industry Corporation of China,Chengdu 611730,,China,; 2.Aerospace Information Research Institute,Chinese Academy of Sciences,,Beijing 100190,,China
Abstract: Synthetic aperture radar(SAR) is widely used in military-reconnaissance, ocean monitoring, disaster emergency assessment and other applications. Among them, ship detection and classification is an important ocean application. Due to the special imaging mechanism, the target features in SAR images are always changeable, which leads to the slow progress of target classification based on SAR images. In recent years, airborne SAR system has developed rapidly due to its real-time and convenience in deployment, and there is a strong demand for ship detection and classification based on airborne SAR images, especially for real-time detection and classification application. Focusing on this issue, a real-time application framework for ship detection and classification based on airborne SAR images is discussed and constructed, and the key processes and methods in engineering application are illustrated. In addition, its development trend combined with the artificial intelligence technology is also prospected.
Key words : airborne;synthetic aperture radar(SAR),;real-time,;ship detection;ship classification

0 引言

    合成孔徑雷達(dá)(Synthetic Aperture Radar,,SAR)作為一種主動(dòng)式微波遙感載荷系統(tǒng),,具有很強(qiáng)的工作環(huán)境適應(yīng)能力,可以在煙塵,、云霧,、陰雨以及夜間等復(fù)雜氣象條件下執(zhí)行觀察任務(wù)。其中,,機(jī)載SAR系統(tǒng)因其使用機(jī)動(dòng)靈活,、分辨率高、實(shí)時(shí)性強(qiáng)等特點(diǎn),,被廣泛應(yīng)用于軍事偵察,、海洋監(jiān)測(cè),、災(zāi)害應(yīng)急評(píng)估等各類應(yīng)用[1-3],表現(xiàn)出極大的應(yīng)用價(jià)值和潛力,。

    海面船只監(jiān)視是世界沿海各國(guó)的傳統(tǒng)任務(wù),,主要涉及船只搜救、非法捕撈和傾倒的監(jiān)視,、軍船活動(dòng)監(jiān)測(cè)等,。傳統(tǒng)的監(jiān)測(cè)手段投入大、覆蓋范圍小,,很難滿足日益加劇的海洋生產(chǎn),、運(yùn)輸和軍事活動(dòng)的監(jiān)查需求。因此,,基于遙感的海面船只監(jiān)測(cè)手段成為各國(guó)研究和發(fā)展的重點(diǎn)。其中,,SAR遙感技術(shù)因其抗云雨和晝夜工作的能力受到廣泛關(guān)注,,特別是基于SAR圖像的船只目標(biāo)檢測(cè)、分類在SAR海洋應(yīng)用中備受關(guān)注,,而針對(duì)實(shí)時(shí)應(yīng)用的海面船只檢測(cè),、分類最具應(yīng)用價(jià)值[4-5]。針對(duì)該應(yīng)用需求,,近二三十年來(lái),,國(guó)內(nèi)外不同研究機(jī)構(gòu)開(kāi)展了大量研究、推進(jìn)工作,,特別是隨著深度學(xué)習(xí)人工智能技術(shù)的興起,,基于深度學(xué)習(xí)方法的SAR艦只目標(biāo)分類問(wèn)題日益受到國(guó)內(nèi)外廣泛關(guān)注[6-10]。為了支撐方法研究,,近年來(lái)人們先后構(gòu)建和公布了多個(gè)SAR船只樣本數(shù)據(jù)集,,包括:SSDD、Air-SAR-Ship,、OpenSARShip等,,極大地推動(dòng)了應(yīng)用研究工作的開(kāi)展。但是,,截至目前,,實(shí)用性的SAR船只實(shí)時(shí)檢測(cè)識(shí)別系統(tǒng)仍停留在目標(biāo)檢測(cè)階段。究其原因,,大致如下:(1)現(xiàn)有數(shù)據(jù)集均為星載SAR數(shù)據(jù)集,,缺乏高質(zhì)量、高分辨率的SAR目標(biāo)數(shù)據(jù)集,,很難支撐分類應(yīng)用需求,;(2)SAR目標(biāo)特征隨系統(tǒng)入射角,、方位角、分辨率,、信噪比及參數(shù)差異等變化較大,,小樣本量的目標(biāo)圖像很難支撐實(shí)際的分類應(yīng)用;(3)對(duì)于實(shí)時(shí)應(yīng)用的機(jī)載SAR系統(tǒng),,星載SAR目標(biāo)樣本缺乏系統(tǒng)普適性,。

    為此,本文針對(duì)SAR海面船只目標(biāo)實(shí)時(shí)分類應(yīng)用中所面臨的上述問(wèn)題,,從實(shí)際的工程應(yīng)用出發(fā),,從系統(tǒng)方案設(shè)計(jì)入手,探討了基于SAR圖像的海面船只實(shí)時(shí)分類的構(gòu)建方案與實(shí)現(xiàn)方法,。




本文詳細(xì)內(nèi)容請(qǐng)下載:http://forexkbc.com/resource/share/2000004015,。




作者信息:

繆煒星1,王小龍2,,李志勇2,,楊  松2,張  欣2,,季昕寧2

(1.中航(成都)無(wú)人機(jī)系統(tǒng)股份有限公司,,四川 成都611730;2.中國(guó)科學(xué)院 空天信息創(chuàng)新研究院,,北京100190)




wd.jpg

此內(nèi)容為AET網(wǎng)站原創(chuàng),,未經(jīng)授權(quán)禁止轉(zhuǎn)載。