基于IDL的无人机遥感影像拼接研究与实现
|
摘要
目前,卫星遥感技术已经在水灾、旱灾、森林火灾、地震、滑坡、泥石流等灾害的调查和灾损评估中得到广泛应用。由于卫星遥感的时间分辨率和空间分辨率的限制,对于灾害的事后分析有一定得价值,但基本不能满足减灾救灾灾情调查与评估的时效性要求。但是无人机遥感技术,可以弥补卫星遥感、航空遥感时效性差的不足。
交互式数据语言IDL是一个面向数组的语言,因此广泛应用于数字遥感图像处理。经过30多年的发展,IDL语言进一步强化了图像处理、分析和显示功能,深受广大学者的欢迎。
SIFT(尺度不变特征转换,Scale Invariant Feature Transform)是目前广泛应用且强健的特征点提取与特征点描述的方法。它最初作为一种关键点的特征提出来的,其主要思想是在尺度空间寻找极值点,提取位置,尺度,旋转不变量;广泛应用于影像融合、影像镶嵌等领域。
本文将重点研究利用IDL语言,实现无人机遥感影像无地面控制点的自动拼接。首先对遥感影像进行镜头畸变和姿态参数同步纠正,然后采用SIFT算法进行自动化特征点描述,提取影像间的特征点,并进行特征点匹配,最终根据最小二乘法原理实现区域全景图的拼接。
研究结果表明,采用SIFT算法提取特征点实现基于普通非量测相机无人机遥感系统获取的影像快速自动拼接在技术上是可行的,满足减灾救灾工作的突发性要求。
At present, the satellite remote sensing technology has been widely applied in floods, droughts, forest fires, earthquakes, landslides, mudslides and other disasters and disaster damage assessment. Because of the limitation of temporal resolution and spatial resolution, satellite remote sensing images have a certain value for post-disaster analysis, and do not meet the timeliness requirements of disaster survey and assessment. UAV remote sensing technology can remedy detects of satellite remote sensing, aerial remote sensing.
Interactive Data Language is an array-oriented language, it is widely used in remote sensing digital image processing. After more than 30 years development, the IDL strengthened the imagery processing, the analysis and the demonstration function. Many scholars want to use it.
SIFT(Scale Invariant Feature Transform)is a widely used and robust feature points extraction and described method. It is initially as a key feature to be proposed. The main idea is to find extreme value in the scale space, extracting location, scale and rotation invariant. It is widely used in image fusion, image mosaics and other fields.
This article will focus on the realization of UAV remote sensing image mosaic without ground control points. Firstly, image must be processed by lens distortion rectification and posture parameters simultaneous rectification; secondly, key points will be extracted and matched by SIFT algorithm; finally, image will be spliced by least square method.
The results of this study show that it is feasible for image rapid automatic stitching by SIFT key point extraction. It also meets the disaster reduction and relief requirements.
交互式数据语言IDL是一个面向数组的语言,因此广泛应用于数字遥感图像处理。经过30多年的发展,IDL语言进一步强化了图像处理、分析和显示功能,深受广大学者的欢迎。
SIFT(尺度不变特征转换,Scale Invariant Feature Transform)是目前广泛应用且强健的特征点提取与特征点描述的方法。它最初作为一种关键点的特征提出来的,其主要思想是在尺度空间寻找极值点,提取位置,尺度,旋转不变量;广泛应用于影像融合、影像镶嵌等领域。
本文将重点研究利用IDL语言,实现无人机遥感影像无地面控制点的自动拼接。首先对遥感影像进行镜头畸变和姿态参数同步纠正,然后采用SIFT算法进行自动化特征点描述,提取影像间的特征点,并进行特征点匹配,最终根据最小二乘法原理实现区域全景图的拼接。
研究结果表明,采用SIFT算法提取特征点实现基于普通非量测相机无人机遥感系统获取的影像快速自动拼接在技术上是可行的,满足减灾救灾工作的突发性要求。
At present, the satellite remote sensing technology has been widely applied in floods, droughts, forest fires, earthquakes, landslides, mudslides and other disasters and disaster damage assessment. Because of the limitation of temporal resolution and spatial resolution, satellite remote sensing images have a certain value for post-disaster analysis, and do not meet the timeliness requirements of disaster survey and assessment. UAV remote sensing technology can remedy detects of satellite remote sensing, aerial remote sensing.
Interactive Data Language is an array-oriented language, it is widely used in remote sensing digital image processing. After more than 30 years development, the IDL strengthened the imagery processing, the analysis and the demonstration function. Many scholars want to use it.
SIFT(Scale Invariant Feature Transform)is a widely used and robust feature points extraction and described method. It is initially as a key feature to be proposed. The main idea is to find extreme value in the scale space, extracting location, scale and rotation invariant. It is widely used in image fusion, image mosaics and other fields.
This article will focus on the realization of UAV remote sensing image mosaic without ground control points. Firstly, image must be processed by lens distortion rectification and posture parameters simultaneous rectification; secondly, key points will be extracted and matched by SIFT algorithm; finally, image will be spliced by least square method.
The results of this study show that it is feasible for image rapid automatic stitching by SIFT key point extraction. It also meets the disaster reduction and relief requirements.
引文
[1]马宗晋,高庆华,陈建英,等.减灾事业的发展和综合减灾.自然灾害学报(2007). 16(1): 1-6.
[2]申旭辉,单新建,吴云.地震遥感应用趋势与中国地震卫星发展框架.国际地震动态(2007).8: 93-100.
[3]黄小雪,程香菊,罗麟.遥感技术在灾害监测中的应用.四川环境(2004).23(6): 102-106.
[4]龚建华,,赵忠明.四川汶川地震应急无人机遥感信息获取与应用.城市发展研究(2008).15(3): 30-32.
[5]洪宇,龚建华,胡社荣,等.无人机遥感影像获取及后续处理探讨.遥感技术与应用(2008).23(4): 462-467.
[6]晏磊,吕书强,赵红颖,等.无人机航空遥感系统关键技术研究.武汉大学学报(工学版)(2004).37(6): 85-88.
[7]姬渊,秦志远,王秉杰,等.小型无人机遥感平台在摄影测量中的应用研究.测绘技术装备(2008).10(1): 46-48.
[8]马轮基,马瑞升,林宗桂,等.微型无人机遥感应用初探.广西气象(2005).26(增刊): 180-181.
[9]吕书强,晏磊,张兵,等.无人机遥感系统的集成于飞行试验研究.测绘科学(2007).32(1):84-87
[10]崔红霞,林宗坚,孙杰.无人机遥感监测系统研究.测绘通报(2005). (5): 11-14.
[11]姜挺,江刚武.基于小波变换的分层影像匹配.测绘学报(2004).33(3): 243-248.
[12]徐建斌,洪文,吴一戎.基于遗传算法的遥感影像匹配定位的研究.测试技术学报(2004).18(4): 351-354.
[13]熊兴华,钱曾波,王任亨.遗传算法与最小二乘法相结合的遥感图像子像素匹配.测绘学报(2001).30(1): 54-59
[14]徐建斌,洪文,吴一戎.基于小波变换和遗传算法的遥感影像匹配方法的研究.电子与信息学报(2005).27(2): 283-285.
[15]安如,王慧麟.基于影像尺度空间表达与鲁棒Hausdorff距离的快速角点特征匹配方法.测绘学报(2005).34(2): 101-107.
[16]王宇宙,汪国平.基于局部仿射不变特征的宽基线影像匹配.计算机应用(2006).26(5): 1001-1004.
[17]朱庆,吴波,赵杰.基于自适应三角形约束的可靠影像匹配方法.计算机学报(2005).28(10): 1734-1738.
[18]熊兴华,陈鹰,钱曾波.一种快速、高精度和稳健的影像匹配算法.测绘学报(2005).34(1): 40-45.
[19] Lowe, D. G. Distinctive image feature from scale invariant key-points. International Journal of Computer Vision (2004).60(2): 91-110.
[20] K.Mikolajczyk, C. Schmid. Scale & affine invariant interest point detectors. International Journal of Computer Vision (2004).60(1): 63-86.
[21]李晓明,郑链,胡占义.基于SIFT特征的遥感影像自动配准.遥感学报(2006).10(6): 885-892.
[22]霍春雷,周志鑫,刘青山,等.基于SIFT特征和广义紧互对原型对距离的遥感图像配准方法.遥感技术与应用(2007).22(4): 524-530.
[23]韩培友. IDL可视化分析与应用.陕西:西北工业大学出版社,2006
[24]路文海. IDL语言处理HDF格式遥感数据的研究.海洋信息(2006).3: 6-7.
[25]张剑清,潘励,王树根.武汉:武汉大学出版社,2003
[26]马瑞升,孙涵,林宗桂.微型无人机遥感影像的纠偏与定位.南京气象学院学报(2005).28(5): 632-639.
[27]崔红霞,孙杰,林宗坚,等.非量测数码相机的畸变差检测研究.测绘科学(2005).30(1): 105-107.
[28]程效军,胡敏捷.数字相机畸变差的检测.测绘学报(2002).31(Sup): 113-117.
[29]卫征,方俊永,张兵.非量测相机镜头光学畸变的改正.光学技术(2007).33(6): 885-888.
[30]李天子,郭辉.非量测数码相机的影像纠正.测绘通报(2006). (10): 59-61.
[31]张朝伟,周焰,吴思励,等.基于SIFT特征匹配的监控图像自动拼接.计算机应用(2008).28(1): 191-194.
[32] Zitov, B. and J. Flusser (2003). Image registration methods: a survey. Image and Vision Computing 21(11): 977-1000.
[33]杨林,张卡,盛业华,等.数字图像的两种影像拼接方法及其比较.计算机应用研究(2005). (5): 135-136.
[34] Baumberg A. Reliable feature matching across widely separated views, Proc. CVPR, (2000):774-781.
[35]陈志雄.基于图像配准的SIFT算法研究与实现:[硕士学位论文].湖北:武汉理工大学,2008
[36] Goshtasby A. A symbolically-assisted approach to digital image registration with application in computer vision. Ph.D.dissertation, Michigan State University, 1983.
[37] Lowe, D. G. Object recognition from Local scale invariant features. International Journal of Computer Vision: (1999). 1150-1157.
[38] Yan Ke,Rahul Sukthankar. PCA-SIFT: A More Distinctive Representation for Local Image Descriptors. 2004
[39] C. Harris and M. Stephens, A combined corner and edge detector, In Alvey Vision Conference, 1988. 147-151.
[40] Brown,M Lowe, D.G. Invariant feature from interest point groups. British Machine Vision Conference (2002).656-665.
[41]王国美,陈孝威. SIFT特征匹配算法研究.盐城工学院学报(自然科学版) (2007).20(2): 1-6.
[42] Koenderink, J. J. The structure of images. Biological Cybernetics (1984).50(5): 363-396.
[43] Lindeberg, T. Scale-space theory: a basic tool for analysing structures at different scale. Journal of Applied Statistics (1999).60(2): 224-270.
[44] Babaud J, Witkin A P, Baudin Metal. Uniqueness of the Gaussian kernel for scale-space filtering. Transactions on Pattern Analysis and Machine Intelligence (1996).8(1): 26-33.
[45] Dufournaud, Cordelia Schmid, Radu Horaud. Image matching with scale adiustment. Computer Vision and Image Understanding. (2004).93(2): 175-194
[46]李玲玲,李翠华,曾晓明,等.基于Harris-Affine和SIFT特征匹配的图像自动配准.华中科技大学学报(自然科学版) (2008).36(8): 13-16
[47]吴锐航.基于SIFT特征的图像检索技术研究:[硕士学位论文].福建:厦门大学,2007
[48]张朝伟,周焰,王耀康,等.基于SIFT特征跟踪匹配的视频拼接方法.计算机工程与应用(2008).44(10): 169-173.
[49]宰小涛,赵宇明.基于SIFT特征描述子的立体匹配算法.微计算机信息(2007).23(8):285-288.
[50]陈信华SIFT特征匹配在无人机低空遥感影像处理中的应用.地矿测绘(2008).24(2): 10-12.
[2]申旭辉,单新建,吴云.地震遥感应用趋势与中国地震卫星发展框架.国际地震动态(2007).8: 93-100.
[3]黄小雪,程香菊,罗麟.遥感技术在灾害监测中的应用.四川环境(2004).23(6): 102-106.
[4]龚建华,,赵忠明.四川汶川地震应急无人机遥感信息获取与应用.城市发展研究(2008).15(3): 30-32.
[5]洪宇,龚建华,胡社荣,等.无人机遥感影像获取及后续处理探讨.遥感技术与应用(2008).23(4): 462-467.
[6]晏磊,吕书强,赵红颖,等.无人机航空遥感系统关键技术研究.武汉大学学报(工学版)(2004).37(6): 85-88.
[7]姬渊,秦志远,王秉杰,等.小型无人机遥感平台在摄影测量中的应用研究.测绘技术装备(2008).10(1): 46-48.
[8]马轮基,马瑞升,林宗桂,等.微型无人机遥感应用初探.广西气象(2005).26(增刊): 180-181.
[9]吕书强,晏磊,张兵,等.无人机遥感系统的集成于飞行试验研究.测绘科学(2007).32(1):84-87
[10]崔红霞,林宗坚,孙杰.无人机遥感监测系统研究.测绘通报(2005). (5): 11-14.
[11]姜挺,江刚武.基于小波变换的分层影像匹配.测绘学报(2004).33(3): 243-248.
[12]徐建斌,洪文,吴一戎.基于遗传算法的遥感影像匹配定位的研究.测试技术学报(2004).18(4): 351-354.
[13]熊兴华,钱曾波,王任亨.遗传算法与最小二乘法相结合的遥感图像子像素匹配.测绘学报(2001).30(1): 54-59
[14]徐建斌,洪文,吴一戎.基于小波变换和遗传算法的遥感影像匹配方法的研究.电子与信息学报(2005).27(2): 283-285.
[15]安如,王慧麟.基于影像尺度空间表达与鲁棒Hausdorff距离的快速角点特征匹配方法.测绘学报(2005).34(2): 101-107.
[16]王宇宙,汪国平.基于局部仿射不变特征的宽基线影像匹配.计算机应用(2006).26(5): 1001-1004.
[17]朱庆,吴波,赵杰.基于自适应三角形约束的可靠影像匹配方法.计算机学报(2005).28(10): 1734-1738.
[18]熊兴华,陈鹰,钱曾波.一种快速、高精度和稳健的影像匹配算法.测绘学报(2005).34(1): 40-45.
[19] Lowe, D. G. Distinctive image feature from scale invariant key-points. International Journal of Computer Vision (2004).60(2): 91-110.
[20] K.Mikolajczyk, C. Schmid. Scale & affine invariant interest point detectors. International Journal of Computer Vision (2004).60(1): 63-86.
[21]李晓明,郑链,胡占义.基于SIFT特征的遥感影像自动配准.遥感学报(2006).10(6): 885-892.
[22]霍春雷,周志鑫,刘青山,等.基于SIFT特征和广义紧互对原型对距离的遥感图像配准方法.遥感技术与应用(2007).22(4): 524-530.
[23]韩培友. IDL可视化分析与应用.陕西:西北工业大学出版社,2006
[24]路文海. IDL语言处理HDF格式遥感数据的研究.海洋信息(2006).3: 6-7.
[25]张剑清,潘励,王树根.武汉:武汉大学出版社,2003
[26]马瑞升,孙涵,林宗桂.微型无人机遥感影像的纠偏与定位.南京气象学院学报(2005).28(5): 632-639.
[27]崔红霞,孙杰,林宗坚,等.非量测数码相机的畸变差检测研究.测绘科学(2005).30(1): 105-107.
[28]程效军,胡敏捷.数字相机畸变差的检测.测绘学报(2002).31(Sup): 113-117.
[29]卫征,方俊永,张兵.非量测相机镜头光学畸变的改正.光学技术(2007).33(6): 885-888.
[30]李天子,郭辉.非量测数码相机的影像纠正.测绘通报(2006). (10): 59-61.
[31]张朝伟,周焰,吴思励,等.基于SIFT特征匹配的监控图像自动拼接.计算机应用(2008).28(1): 191-194.
[32] Zitov, B. and J. Flusser (2003). Image registration methods: a survey. Image and Vision Computing 21(11): 977-1000.
[33]杨林,张卡,盛业华,等.数字图像的两种影像拼接方法及其比较.计算机应用研究(2005). (5): 135-136.
[34] Baumberg A. Reliable feature matching across widely separated views, Proc. CVPR, (2000):774-781.
[35]陈志雄.基于图像配准的SIFT算法研究与实现:[硕士学位论文].湖北:武汉理工大学,2008
[36] Goshtasby A. A symbolically-assisted approach to digital image registration with application in computer vision. Ph.D.dissertation, Michigan State University, 1983.
[37] Lowe, D. G. Object recognition from Local scale invariant features. International Journal of Computer Vision: (1999). 1150-1157.
[38] Yan Ke,Rahul Sukthankar. PCA-SIFT: A More Distinctive Representation for Local Image Descriptors. 2004
[39] C. Harris and M. Stephens, A combined corner and edge detector, In Alvey Vision Conference, 1988. 147-151.
[40] Brown,M Lowe, D.G. Invariant feature from interest point groups. British Machine Vision Conference (2002).656-665.
[41]王国美,陈孝威. SIFT特征匹配算法研究.盐城工学院学报(自然科学版) (2007).20(2): 1-6.
[42] Koenderink, J. J. The structure of images. Biological Cybernetics (1984).50(5): 363-396.
[43] Lindeberg, T. Scale-space theory: a basic tool for analysing structures at different scale. Journal of Applied Statistics (1999).60(2): 224-270.
[44] Babaud J, Witkin A P, Baudin Metal. Uniqueness of the Gaussian kernel for scale-space filtering. Transactions on Pattern Analysis and Machine Intelligence (1996).8(1): 26-33.
[45] Dufournaud, Cordelia Schmid, Radu Horaud. Image matching with scale adiustment. Computer Vision and Image Understanding. (2004).93(2): 175-194
[46]李玲玲,李翠华,曾晓明,等.基于Harris-Affine和SIFT特征匹配的图像自动配准.华中科技大学学报(自然科学版) (2008).36(8): 13-16
[47]吴锐航.基于SIFT特征的图像检索技术研究:[硕士学位论文].福建:厦门大学,2007
[48]张朝伟,周焰,王耀康,等.基于SIFT特征跟踪匹配的视频拼接方法.计算机工程与应用(2008).44(10): 169-173.
[49]宰小涛,赵宇明.基于SIFT特征描述子的立体匹配算法.微计算机信息(2007).23(8):285-288.
[50]陈信华SIFT特征匹配在无人机低空遥感影像处理中的应用.地矿测绘(2008).24(2): 10-12.