摄影测量技术在交通事故现场快速勘查中的应用
|
摘要
交通事故处理过程大致分为三个阶段：一是事故现场勘查取证；二是事故原因和过程分析；三是根据分析结果进行事故责任认定。在上述三个阶段中，第一阶段工作是在交通事故现场完成的，这一阶段持续时间的长短直接决定了事故对交通影响的程度。目前，我国交警部门处理事故现场的常用方法是：先封锁事故地点的部分或全部车道，工作人员通过皮尺测量、照相、目测等方法对事故现场进行勘查，对现场情况进行详尽的调查取证之后开放交通。这种处理方法能够对交通事故的处理取得足够的材料，但也存在明显的缺陷。首先对交通的影响较大，当事故较严重时，道路上车辆的排队长度可能达几公里。其次这种方法自动化程度不高，存在着人为影响因素，事故现场资料存储、建档和检索有一定困难，以及现场立体再现和现场恢复困难，测量数据精度较低等方面的问题，另外受环境、天气和时间制约较大。因此，对传统交通事故现场勘查手段的改进势在必行。
目前，用于交通事故现场勘查的主要测量方法有皮尺、全站仪、激光、GPS、摄影测量五种测量方法。其中摄影测量方法又分为立体照相机测量法和普通照相机测量法。
1933年，瑞士的苏黎世警局首次采用了立体照相机进行现场勘查，随后的德国、意大利、英国等欧洲警方也有一部分采用。目前，应用立体照相机最为广泛的国家为日本。立体照相机的主要优点是成像品质较佳，还原现场的能力迅速又精确；缺点是立体照相机的成本较高、设备操作及维护不易，非经特别训练，无法顺利操作，需有专人保管使用。
在交通事故勘查中使用普通照相机立体测量，对执法机关来说，仍然是一
个相对较新并正在测试的方法。由于目前计算机技术、数码相机的飞速发展，使得硬件的成本相应的降低，再加上普通相机立体测量软件的不断成熟，使得美国、加拿大等国家对在交通事故现场勘查中使用这种方法的兴趣也越来越浓。目前，美国的华盛顿、亚利桑那、犹他州、内华达、加利福尼亚、俄勒冈、檀香山等地区的执法部门正在积极测试这种方法。测试的结果表明，普通照相机立体测量法相对于目前美国正在使用的全站仪测量法来说，是一种相对操作简便、简单易学、节约成本、精度较高的测量方法。同时摄影测量方法又具有全站仪方法所不具有的独特优点：节约现场勘查时间、随时分析照片中任何想要测量的数据，包括现场非常容易丢失的轮胎痕迹、血迹等信息。普通照相机立体测量很有可能取代其他全站仪而成为今后交通事故勘查中的主要方法，而其他测量方式只是在普通照相机立体测量方法有困难时的有益补充。
本论文基于国家自然科学基金重点项目“城市路网动态交通管理与控制关键理论及模拟技术研究”子项目“交通事故现场快速处理技术研究”，主要针对2008年北京奥运会，期望研究一种能够快速高效处理交通事故现场的新技术——摄影测量技术。它以数字图像和摄影测量的基本原理为基础，应用图像预处理技术、相机内外参数标定技术及立体匹配算法从所摄取图像中准确获取空间点的三维坐标。
本论文研究的主要内容是：第一章,在分析了目前国内外交通事故现场处理方法现状和应用情况的基础上，针对目前具有较强应用前景的摄影测量技术，提出了较为完整的交通事故现场快速摄影测量研究方案。第二章,摄影测量的图像处理技术。主要是介绍对所拍摄图像对进行处理的一些方法，包括图像的增强、放大、平滑等技术，并介绍了用于控制点(或特征点)提取的点特征提取算子以及线特征提取算子。第三章，摄影测量的基本理论。利用定位后的控制点对照相机内外参数进行标定，建立图像平面二维坐标和空间三维坐标之间的转换模型。第四章，立体匹配算法。主要是对如何快速准确寻找空间上同一点在
两幅图像上的对应点进行研究。首先提出对立体匹配非常有帮助的极线约束，然后介绍了灰度匹配和特征匹配两种匹配方法及相应的匹配策略。最后，基于以上两种匹配方法的不足，提出了一种基于颜色分割的全局整体匹配标准。第五章，实验部分。首先对张正友提出的一种由较为灵活的基于未知方向平面图像的二维相机定标算法进行介绍，在此基础上通过软件编程验证该定标算法的有效性。
交通事故现场的快速勘查在缓解交通压力、减轻交通拥挤等方面具有非常重要的作用。它不仅在一定程度上代表着我国交通事故处理的科技水平，同时也是我们国家经济快速发展对交通事故处理的新的需求。因此，摄影测量技术在交通事故现场快速勘查中的应用研究意义重大，应用前景良好。
The process of handling traffic accidents is mainly divided into three stages: The first is investigation and evidence-taking of the traffic accident scene, the second is reason-analyzing and process -analyzing of traffic accident, the third is duty-affirming of traffic accident according to the result of analysis. In above three stages, the first stage work completes in traffic accident scene. The duration time of the the first stage decides the degree of the traffic affected directly. Currently, the general method of our police departments in investigation of traffic accidents is: firstly, close off part or all driveways of the traffic accident scene, then inspectors investigate the traffic accident scene by tape measure, photos and eye-measure and etc. Lastly, resume the traffic after taking the detail and entire evidence. This kind of method of handling the traffic accident can get enough material, but also exits obvious limitations. The first is obvious influence of the traffic, if the traffic accident is serious, the length of the car in road may reach a few kilometers. The next in order is that the automatization degree of this kind of method is not high, affected by the artificial factor, having certain difficulty in storing ,creating and searching files of traffic accident scene, also having difficulty in reconstruction and reappear stereo scene, having the problem of the low accuracy of measured data. At the same, the environment, weather and time can affect the result of measure. Therefore, improving the means of the traditional traffic accident investigation is necessary.
Currently, the main measure methods used in traffic accident scene investigation include: Tape Measure, Total Stations, Laser Units, Global Postitoning
System, Photogrammetry. Among them Photogrammetry is divided into stereo camera photogrammetry and common camera Photogrammetry.
In 1933, Swiss police station adopted stereo camera Photogrammetry in investigation of traffic accident scene firstly, then European police stations such as Germany, Italy, England etc. also had a part adoption. Nowadays, the nation of applying stereo camera Photogrammetry most extensively is Japan. The main advantages of using stereo camera Photogrammetry are: better image quality, accurate and quick ability of resuming traffic accident scene. The weaknesses of using stereo camera Photogrammetry are: higher cost, difficult equipments operation and maintenance, it can’t operate smoothly if not through special training, and it also need specially person take care of and use.
The use of photogrammetry by law enforcement agencies for the investigation of traffic accidents is still a relatively new and under-test method. Because of rapid development of computer technique and digital cameras, it makes computer hardware cost drop fast, and also because common camera Photogrammetry software become more and more mature, the country such as America ,Canada and etc. have more and more interest in using common camera Photogrammetry in investigation of traffic accident scene. At present, law enforcement agencies for the investigation of traffic accidents such as Washington, Arizona, Utah, Nevada, California, Oregon, Honolulu are testing the method actively. The result of the test shows: as to the Total Stations method used these days in US, the common camera Photogrammetry method is relatively simple to operate, easy to learn, lower cost, and higher accuracy. At the same time, the common camera Photogrammetry method has its special advantages: investigation time-saving, analysis of measure data needed in photo at any time, include tyre mark, blood block and etc. which are
easy to lose. The common camera Photogrammetry method will be potentially replace the Total Stations method in the future and become the main method used in investigation of traffic accident scene, other methods are just used when the common camera Photogrammetry method have some difficulty.
Based on the sub-item" Research on Traffic Accident Scene Fast Investigation Technique "of
目前，用于交通事故现场勘查的主要测量方法有皮尺、全站仪、激光、GPS、摄影测量五种测量方法。其中摄影测量方法又分为立体照相机测量法和普通照相机测量法。
1933年，瑞士的苏黎世警局首次采用了立体照相机进行现场勘查，随后的德国、意大利、英国等欧洲警方也有一部分采用。目前，应用立体照相机最为广泛的国家为日本。立体照相机的主要优点是成像品质较佳，还原现场的能力迅速又精确；缺点是立体照相机的成本较高、设备操作及维护不易，非经特别训练，无法顺利操作，需有专人保管使用。
在交通事故勘查中使用普通照相机立体测量，对执法机关来说，仍然是一
个相对较新并正在测试的方法。由于目前计算机技术、数码相机的飞速发展，使得硬件的成本相应的降低，再加上普通相机立体测量软件的不断成熟，使得美国、加拿大等国家对在交通事故现场勘查中使用这种方法的兴趣也越来越浓。目前，美国的华盛顿、亚利桑那、犹他州、内华达、加利福尼亚、俄勒冈、檀香山等地区的执法部门正在积极测试这种方法。测试的结果表明，普通照相机立体测量法相对于目前美国正在使用的全站仪测量法来说，是一种相对操作简便、简单易学、节约成本、精度较高的测量方法。同时摄影测量方法又具有全站仪方法所不具有的独特优点：节约现场勘查时间、随时分析照片中任何想要测量的数据，包括现场非常容易丢失的轮胎痕迹、血迹等信息。普通照相机立体测量很有可能取代其他全站仪而成为今后交通事故勘查中的主要方法，而其他测量方式只是在普通照相机立体测量方法有困难时的有益补充。
本论文基于国家自然科学基金重点项目“城市路网动态交通管理与控制关键理论及模拟技术研究”子项目“交通事故现场快速处理技术研究”，主要针对2008年北京奥运会，期望研究一种能够快速高效处理交通事故现场的新技术——摄影测量技术。它以数字图像和摄影测量的基本原理为基础，应用图像预处理技术、相机内外参数标定技术及立体匹配算法从所摄取图像中准确获取空间点的三维坐标。
本论文研究的主要内容是：第一章,在分析了目前国内外交通事故现场处理方法现状和应用情况的基础上，针对目前具有较强应用前景的摄影测量技术，提出了较为完整的交通事故现场快速摄影测量研究方案。第二章,摄影测量的图像处理技术。主要是介绍对所拍摄图像对进行处理的一些方法，包括图像的增强、放大、平滑等技术，并介绍了用于控制点(或特征点)提取的点特征提取算子以及线特征提取算子。第三章，摄影测量的基本理论。利用定位后的控制点对照相机内外参数进行标定，建立图像平面二维坐标和空间三维坐标之间的转换模型。第四章，立体匹配算法。主要是对如何快速准确寻找空间上同一点在
两幅图像上的对应点进行研究。首先提出对立体匹配非常有帮助的极线约束，然后介绍了灰度匹配和特征匹配两种匹配方法及相应的匹配策略。最后，基于以上两种匹配方法的不足，提出了一种基于颜色分割的全局整体匹配标准。第五章，实验部分。首先对张正友提出的一种由较为灵活的基于未知方向平面图像的二维相机定标算法进行介绍，在此基础上通过软件编程验证该定标算法的有效性。
交通事故现场的快速勘查在缓解交通压力、减轻交通拥挤等方面具有非常重要的作用。它不仅在一定程度上代表着我国交通事故处理的科技水平，同时也是我们国家经济快速发展对交通事故处理的新的需求。因此，摄影测量技术在交通事故现场快速勘查中的应用研究意义重大，应用前景良好。
The process of handling traffic accidents is mainly divided into three stages: The first is investigation and evidence-taking of the traffic accident scene, the second is reason-analyzing and process -analyzing of traffic accident, the third is duty-affirming of traffic accident according to the result of analysis. In above three stages, the first stage work completes in traffic accident scene. The duration time of the the first stage decides the degree of the traffic affected directly. Currently, the general method of our police departments in investigation of traffic accidents is: firstly, close off part or all driveways of the traffic accident scene, then inspectors investigate the traffic accident scene by tape measure, photos and eye-measure and etc. Lastly, resume the traffic after taking the detail and entire evidence. This kind of method of handling the traffic accident can get enough material, but also exits obvious limitations. The first is obvious influence of the traffic, if the traffic accident is serious, the length of the car in road may reach a few kilometers. The next in order is that the automatization degree of this kind of method is not high, affected by the artificial factor, having certain difficulty in storing ,creating and searching files of traffic accident scene, also having difficulty in reconstruction and reappear stereo scene, having the problem of the low accuracy of measured data. At the same, the environment, weather and time can affect the result of measure. Therefore, improving the means of the traditional traffic accident investigation is necessary.
Currently, the main measure methods used in traffic accident scene investigation include: Tape Measure, Total Stations, Laser Units, Global Postitoning
System, Photogrammetry. Among them Photogrammetry is divided into stereo camera photogrammetry and common camera Photogrammetry.
In 1933, Swiss police station adopted stereo camera Photogrammetry in investigation of traffic accident scene firstly, then European police stations such as Germany, Italy, England etc. also had a part adoption. Nowadays, the nation of applying stereo camera Photogrammetry most extensively is Japan. The main advantages of using stereo camera Photogrammetry are: better image quality, accurate and quick ability of resuming traffic accident scene. The weaknesses of using stereo camera Photogrammetry are: higher cost, difficult equipments operation and maintenance, it can’t operate smoothly if not through special training, and it also need specially person take care of and use.
The use of photogrammetry by law enforcement agencies for the investigation of traffic accidents is still a relatively new and under-test method. Because of rapid development of computer technique and digital cameras, it makes computer hardware cost drop fast, and also because common camera Photogrammetry software become more and more mature, the country such as America ,Canada and etc. have more and more interest in using common camera Photogrammetry in investigation of traffic accident scene. At present, law enforcement agencies for the investigation of traffic accidents such as Washington, Arizona, Utah, Nevada, California, Oregon, Honolulu are testing the method actively. The result of the test shows: as to the Total Stations method used these days in US, the common camera Photogrammetry method is relatively simple to operate, easy to learn, lower cost, and higher accuracy. At the same time, the common camera Photogrammetry method has its special advantages: investigation time-saving, analysis of measure data needed in photo at any time, include tyre mark, blood block and etc. which are
easy to lose. The common camera Photogrammetry method will be potentially replace the Total Stations method in the future and become the main method used in investigation of traffic accident scene, other methods are just used when the common camera Photogrammetry method have some difficulty.
Based on the sub-item" Research on Traffic Accident Scene Fast Investigation Technique "of
引文
李江 等编著 交通工程学,人民交通出版社,2002.7
李江、江国宪、李洪才 照相测量技术在事故再现中的应用，中国公路学报，1995，增1
朱艳秋、李江、王强. 道路交通事故处理的技术及发展. 公路交通科技，1999:66-68
Olga Scott A. Cooner, P.E. and Kevin N. Balke, Ph.D., P.E.，Use of Photogrammetry for Investigation of Traffic Incident Scenes，Technical Report，TX-99/4907-2，October 2000
Veksler，Stereo Matching by Compact Windows via Minimum Ratio Cycle，IEEE International Conference on Computer Vision (July 2001)
Hai Tao and Harpreet S. Sawhney，Global Matching Criterion and Color Segmentation Based Stereo in Proc. Workshop on the Application of Computer Vision (WACV2000), pp. 246-253, December 2000
L. Di Stefano, M. Marchionni, S. Mattoccia, G. Neri，A Fast Area-Based Stereo Matching Algorithm，15th IAPR /CIPPRS International Conference on Vision Interface (VI'2002),May 27-29, 2002, Calgary, Canada
Zhengyou Zhang，A Flexible New Technique for Camera Calibration，Technical Report，MSR-TR-98-71，December 2, 1998
O. Faugeras and G. Toscani. “The Calibration Problem for Stereo”, Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, pages 15–20, Miami Beach, FL, June 1986.
J. Weng, P. Cohen, and M. Herniou. Camera calibration with distortion models and accuracy evaluation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(10):965–980, Oct. 1992.
W. Faig. Calibration of close-range photogrammetry systems: Mathematical formulation. Photogrammetric Engineering and Remote Sensing, 41(12):1479–1486, 1975.
张祖勋 张剑清 编著，数字摄影测量学，武汉大学出版社，1997.1
马颂德 张正友著，计算机视觉：计算理论与算法基础，科学出版社，1998.1章毓晋. 图像理解与计算机视觉. 清华大学出版社, 2000.8
章毓晋编著，图像处理和分析基础，章毓晋编著2002.6
www.sanliantec.com
www.gildsoft.com
www.chinacvic.com
周坚清，事故再现中计算机图像处理技术的应用，吉林工业大学硕士研究生毕业论文，1996年1月
艾海周，数字图象处理，清华大学计算机科学与技术系，人机交互与媒体集成研究所(电子版)，2000年1月
孙伟杰，计算机视觉理论与方法研究，吉林工业大学硕士研究生毕业论文，2000年2月
郑黎黎，基于计算机视觉的单目测量方法，吉林工业大学硕士研究生毕业论文，2002年2月
徐飞, 施晓红等编著，MATLAB应用图像处理，西安电子科技大学出版社，2002
董长虹编著，Matlab图像处理与应用，国防工业出版社，2004.1
郑南宁著，计算机视觉与模式识别，国防工业出版社，1998
[美]Milan Sonka,[美]Vaclav Hlavac,[美]Roger Boyle，Image processing, analysis, and machine vision:Second edition，人民邮电出版社 2002.1
Milan Sonka, Vaclav Hlavac, Roger Boyle著，图像处理、分析与机器视觉，人民邮电出版社，2003
www.shapecapture.com
29． www.photomodeler.com
30.苏志强编著，交通事故侦查理论与实务，一成印刷有限公司，19997年10月初版。
李江、江国宪、李洪才 照相测量技术在事故再现中的应用，中国公路学报，1995，增1
朱艳秋、李江、王强. 道路交通事故处理的技术及发展. 公路交通科技，1999:66-68
Olga Scott A. Cooner, P.E. and Kevin N. Balke, Ph.D., P.E.，Use of Photogrammetry for Investigation of Traffic Incident Scenes，Technical Report，TX-99/4907-2，October 2000
Veksler，Stereo Matching by Compact Windows via Minimum Ratio Cycle，IEEE International Conference on Computer Vision (July 2001)
Hai Tao and Harpreet S. Sawhney，Global Matching Criterion and Color Segmentation Based Stereo in Proc. Workshop on the Application of Computer Vision (WACV2000), pp. 246-253, December 2000
L. Di Stefano, M. Marchionni, S. Mattoccia, G. Neri，A Fast Area-Based Stereo Matching Algorithm，15th IAPR /CIPPRS International Conference on Vision Interface (VI'2002),May 27-29, 2002, Calgary, Canada
Zhengyou Zhang，A Flexible New Technique for Camera Calibration，Technical Report，MSR-TR-98-71，December 2, 1998
O. Faugeras and G. Toscani. “The Calibration Problem for Stereo”, Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, pages 15–20, Miami Beach, FL, June 1986.
J. Weng, P. Cohen, and M. Herniou. Camera calibration with distortion models and accuracy evaluation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(10):965–980, Oct. 1992.
W. Faig. Calibration of close-range photogrammetry systems: Mathematical formulation. Photogrammetric Engineering and Remote Sensing, 41(12):1479–1486, 1975.
张祖勋 张剑清 编著，数字摄影测量学，武汉大学出版社，1997.1
马颂德 张正友著，计算机视觉：计算理论与算法基础，科学出版社，1998.1章毓晋. 图像理解与计算机视觉. 清华大学出版社, 2000.8
章毓晋编著，图像处理和分析基础，章毓晋编著2002.6
www.sanliantec.com
www.gildsoft.com
www.chinacvic.com
周坚清，事故再现中计算机图像处理技术的应用，吉林工业大学硕士研究生毕业论文，1996年1月
艾海周，数字图象处理，清华大学计算机科学与技术系，人机交互与媒体集成研究所(电子版)，2000年1月
孙伟杰，计算机视觉理论与方法研究，吉林工业大学硕士研究生毕业论文，2000年2月
郑黎黎，基于计算机视觉的单目测量方法，吉林工业大学硕士研究生毕业论文，2002年2月
徐飞, 施晓红等编著，MATLAB应用图像处理，西安电子科技大学出版社，2002
董长虹编著，Matlab图像处理与应用，国防工业出版社，2004.1
郑南宁著，计算机视觉与模式识别，国防工业出版社，1998
[美]Milan Sonka,[美]Vaclav Hlavac,[美]Roger Boyle，Image processing, analysis, and machine vision:Second edition，人民邮电出版社 2002.1
Milan Sonka, Vaclav Hlavac, Roger Boyle著，图像处理、分析与机器视觉，人民邮电出版社，2003
www.shapecapture.com
29． www.photomodeler.com
30.苏志强编著，交通事故侦查理论与实务，一成印刷有限公司，19997年10月初版。