图像处理在木材材积计算中的应用研究
|
摘要
木材材积的测量方法如手工测量查表求取、激光测量等耗费人力物力、成本高、精度低，很难适应当今社会经济的发展和我国各行业对于木材需求量增加的国情。从上世纪90年代初，国内开始从事应用计算机视觉的方法进行木材的在线自动计数和材积测量的研究，取得了一定的成果，但还处于理论研究阶段，没有出现成熟的产品。
基于数字图像处理的木材材积测量技术正是为了克服上述困难应运而生的。木材材积测量技术具有成本低廉，操作简单，处理速度快等优点。木材材积测量系统的研究不仅具有一定的理论意义，同时还具有良好的市场前景。
本课题是采用计算机视觉的方法进行木材材积和数量的测量，主要研究内容是系统的软件部分，包括图像的采集显示、预处理、物体的分割、材积和数量的测量。图像处理中，能否达到满意的处理效果，图像预处理是很关键的一步，它是其它算法的基础，本文根据木材截面的特点，提出了根据颜色空间的提取、改进的图像标记等算法改善目标提取效果，基本上能达到满意的预处理效果。另外根据图像处理的具体特点，文中提出了一种新的基于类圆的粘连体分割方法，当物体间的缝隙在二值化后能部分或全部判定为背景，即粘连的物体在二值化后能有明显的凹陷时，识别率很高，误差几乎可以为零。反之，则误差会上升。最后通过对分割后的图像进行标记处理，计算各个截面的最小直径，找出相对应的木材材积。对图像进一步腐蚀，再做标记，统计标号的个数，即图像中木材的根数。形成木材材积计算的非接触测量系统。
本文在最后进行了试验总结和误差分析，分析了造成系统误差的各种因素，从实际的角度对考察系统的实用性进行分析，指出进一步开发的方向和建议。为木材数量和材积计算软件的商业化打下了良好的基础。
At present, almost all company count log volume one by one by workers or laser light, it is laborious, low effective, also counting by workers may cause high error and high cost. So, these manufactories desire a cheap counter and volume-meter. From the first of 1990s, by using computer vision, studies of this area had been begun. Although a bit progress had been got since then, there is much to do if it is used at spot in factory. This study do some research on theory, the product will have a good future.
Based on digital image process, the log volume-meter technology is born to conquer the drawback of previously methods. It has many merits such as low cost, easy operation, and high efficiency. The research of log volume-meter system not only has the sense of theory, but also has bright market prospect.
In this paper, the author tries to measure the log volume of image by making use of computer vision. The main research includes the following area: image acquisition and display, pre-process, segment of object, measurement of log quantity and volume. During the image process, whether can we get an ideal result depend on the pre-process of image which is the foundation of the whole system. In this paper, the algorithm threshold the object through the specified color space, and mend the image through labeling algorithm. We can get a acceptable result as expected. A new segmentation method based on quasi-circular assumption is presented. This method require a good binary image, if it exist concave, the aggregated objects will be segmented and recognized correctly and the error is lower, otherwise, it may lead to error result. Finally, the software counts the min diameter of each log from the pre-process image. After erosion and labeling of image, it can gain the quantity of the log. This is a non-contact measurement system.
This paper finally presents the experiential result and error analyse. I analyse the practicability of the system from the view of application, and provide suggestions of further development which pave the way for the application of the system.
基于数字图像处理的木材材积测量技术正是为了克服上述困难应运而生的。木材材积测量技术具有成本低廉，操作简单，处理速度快等优点。木材材积测量系统的研究不仅具有一定的理论意义，同时还具有良好的市场前景。
本课题是采用计算机视觉的方法进行木材材积和数量的测量，主要研究内容是系统的软件部分，包括图像的采集显示、预处理、物体的分割、材积和数量的测量。图像处理中，能否达到满意的处理效果，图像预处理是很关键的一步，它是其它算法的基础，本文根据木材截面的特点，提出了根据颜色空间的提取、改进的图像标记等算法改善目标提取效果，基本上能达到满意的预处理效果。另外根据图像处理的具体特点，文中提出了一种新的基于类圆的粘连体分割方法，当物体间的缝隙在二值化后能部分或全部判定为背景，即粘连的物体在二值化后能有明显的凹陷时，识别率很高，误差几乎可以为零。反之，则误差会上升。最后通过对分割后的图像进行标记处理，计算各个截面的最小直径，找出相对应的木材材积。对图像进一步腐蚀，再做标记，统计标号的个数，即图像中木材的根数。形成木材材积计算的非接触测量系统。
本文在最后进行了试验总结和误差分析，分析了造成系统误差的各种因素，从实际的角度对考察系统的实用性进行分析，指出进一步开发的方向和建议。为木材数量和材积计算软件的商业化打下了良好的基础。
At present, almost all company count log volume one by one by workers or laser light, it is laborious, low effective, also counting by workers may cause high error and high cost. So, these manufactories desire a cheap counter and volume-meter. From the first of 1990s, by using computer vision, studies of this area had been begun. Although a bit progress had been got since then, there is much to do if it is used at spot in factory. This study do some research on theory, the product will have a good future.
Based on digital image process, the log volume-meter technology is born to conquer the drawback of previously methods. It has many merits such as low cost, easy operation, and high efficiency. The research of log volume-meter system not only has the sense of theory, but also has bright market prospect.
In this paper, the author tries to measure the log volume of image by making use of computer vision. The main research includes the following area: image acquisition and display, pre-process, segment of object, measurement of log quantity and volume. During the image process, whether can we get an ideal result depend on the pre-process of image which is the foundation of the whole system. In this paper, the algorithm threshold the object through the specified color space, and mend the image through labeling algorithm. We can get a acceptable result as expected. A new segmentation method based on quasi-circular assumption is presented. This method require a good binary image, if it exist concave, the aggregated objects will be segmented and recognized correctly and the error is lower, otherwise, it may lead to error result. Finally, the software counts the min diameter of each log from the pre-process image. After erosion and labeling of image, it can gain the quantity of the log. This is a non-contact measurement system.
This paper finally presents the experiential result and error analyse. I analyse the practicability of the system from the view of application, and provide suggestions of further development which pave the way for the application of the system.
引文
[1] 栾新．原木材积计算机视觉检测系统的研究：(博士论文)．哈尔滨：哈尔滨工业大学，1997．
[2] 王金满等．计算机视觉技术在木材工业科研与生产中的应用．世界林业研究，2001，(3)：49-55．
[3] 黄庆维．木材材积表(2002年修订版)．广州：广东科技出版社，2002．
[4] 韩玉杰等．计算机控制激光扫描识别木段横截面图形的研究．木材工业，2000，(1)：13-14．
[5] Roberto Brunelli, Tomaso Poggio. Face recognition: Feature versus templates. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2000,15(10):1042～1052.
[6] 陈孟英．Visual C++6.0中文入门图解教程．北京希望电子出版社，1999．
[7] 高守传，姚领田．Visual C++实践与提高—数字图像处理与工程应用篇．北京：中国铁道出版社，2006．
[8] Lu X G, Zhou J, Zhang C S. A novel algorithm for rotated Huffman faces detection. In: Proc IEEE Conference on Computer Vision and Pattern Recognition, Hilton Head Island, South Carolina, USA, 2005,760-764.
[9] 王耀南，李树涛，毛建旭．计算机图像处理与识别技术．北京：高等教育出版社，2001．
[10] 贾永红．计算机图像处理与分析．武汉：武汉大学出版社，2002．
[11] 刘智勇，刘迎建．车牌识别(LPR)中的图像提取及分割．中文信息学报，2001，110-169．
[12] Danian Zheng, Yannan Zhao, Jiaxin Wang, An efficient method of license plate location, Pattern Recognition Letters 26(2005)2431-2438.
[13] 杨淑莹．VC++图像处理程序设计．北京：清华大学出版社，2003．
[14] 周长发．精通Visual C++图像编程．北京：电子工业出版社，2006．
[15] Sun Q B, Huang W M, Wu J K. Face detection based on co lo rand local symmetry information. In: Proc Conference Automatic Face and Gesture Recognition, Nara, Japan, 2004.130-135.
[16] Milan Sonka. Vaclav Hlavac. Roger Boyle. Image Processing, Analysis and Machine Vision(Second Edition).USA:Brooks/Cole.
[17] 钟志光，卢勇．Visual C++.net数字图像处理与解析．北京：清华大学山版社，2003．
[18] 易丛琴，李丽，孟传良．基于肤色分割的彩色图像的人脸检测．西安交通大学学报，2003，32(4)：65～68．
[19] 章毓晋．图像处理和分析基础．北京：高等教育出版社，2002．
[20] 吴东晖．智能移动机器人的视觉导航技术：(博士论文)．杭州：浙江大学，2006．
[21] YaoquanYang, Jie Bai, Ruili Tian, et al. A Vehicle License Plate Recognition System based on fixed color collocation. Proc. IEEE the Fourth International Conference on Machine Learning and Cybernetics. Guangzhou. 2005: 18-21.
[22] 龚声蓉等．数字图像处理与分析．北京：清华大学出版社，2006．
[23] Hu Yuchen, Chang Chinchen. Edge Detection Using Block Truncation Coding. International Journal Pattern Recognition and Artificial Intelligence, 2003,17(6): 951-966.
[24] Zhang Y J.A survey on evaluation methods for image segmentation. Pattern Recognition, 2000.29 (8):1335-1346
[25] Chalana V, Kim Y. A Methodology for Evaluation of Boundary Detection on Medical Images. IEEE Trans on Medical Imaging, 1999,16(5):642-652.
[26] 欧珊瑚，王倩丽．Visual C++.NET数字图像处理技术与应用．北京：清华大学出版社，2004．
[27] 栾新，朱铁一．不规则类圆形目标图象识别新策略．中国图象图形学报，1999，4(3)，A：202-206．
[28] 王听，曲波．一种不规则封闭形状的轮廓跟踪算法．南京航空航天大学报，2003，28(4)：58—60．
[29] 韩玉杰等．计算机控制激光扫描识别木段横截面图形的研究．木材工业，1999，(1)：13—14．
[30] 邵峰晶，于忠清．数据挖掘——原理与算法．北京：中国水利水电出版社，2003．
[31] 李国辉，武德峰，付畅俭．多媒体知识及其获取．第十三届全国多媒体技术学术会议论文集，宁波，2004：451～456．
[32] 杨枝灵，王开等．Visual C++数字图像获取处理及实践应用．北京：人民邮电出版社，2003．7382504-573．
[33] 崔屹．图象处理与分析——数学形态学方法及应用．北京：科学出版社，2000．
[34] 刘立波．数学形态学在二值图象处理与分析中的研究及应用：(硕士学位论文)．西安：西安交通大学．2002．
[35] Bruce Eckel. Thinking in C++(Second Edition)Volume One Introduction to Standard C++.MindView, Inc, 2000.
[36] TehChuan Chen, KuoLiang Chung. An Efficient Randomized Algorithm for Detecting Circles Computer Vision and Image Understanding, 2001,83(2).172-191.
[37] Rosenfield A. Computer Vision-Past Present and Future. Information Science, 1999:57-58: 291-243.
[38] 童雀菊等．用图像处理法采集原木形状参数的研究．林业科学，2005，34(3)：87—95．
[39] 冯莉．原木材积计算机视觉检测系统的研究：(硕士学位论文)．哈尔滨：尔北林业大学2002(3)．
[2] 王金满等．计算机视觉技术在木材工业科研与生产中的应用．世界林业研究，2001，(3)：49-55．
[3] 黄庆维．木材材积表(2002年修订版)．广州：广东科技出版社，2002．
[4] 韩玉杰等．计算机控制激光扫描识别木段横截面图形的研究．木材工业，2000，(1)：13-14．
[5] Roberto Brunelli, Tomaso Poggio. Face recognition: Feature versus templates. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2000,15(10):1042～1052.
[6] 陈孟英．Visual C++6.0中文入门图解教程．北京希望电子出版社，1999．
[7] 高守传，姚领田．Visual C++实践与提高—数字图像处理与工程应用篇．北京：中国铁道出版社，2006．
[8] Lu X G, Zhou J, Zhang C S. A novel algorithm for rotated Huffman faces detection. In: Proc IEEE Conference on Computer Vision and Pattern Recognition, Hilton Head Island, South Carolina, USA, 2005,760-764.
[9] 王耀南，李树涛，毛建旭．计算机图像处理与识别技术．北京：高等教育出版社，2001．
[10] 贾永红．计算机图像处理与分析．武汉：武汉大学出版社，2002．
[11] 刘智勇，刘迎建．车牌识别(LPR)中的图像提取及分割．中文信息学报，2001，110-169．
[12] Danian Zheng, Yannan Zhao, Jiaxin Wang, An efficient method of license plate location, Pattern Recognition Letters 26(2005)2431-2438.
[13] 杨淑莹．VC++图像处理程序设计．北京：清华大学出版社，2003．
[14] 周长发．精通Visual C++图像编程．北京：电子工业出版社，2006．
[15] Sun Q B, Huang W M, Wu J K. Face detection based on co lo rand local symmetry information. In: Proc Conference Automatic Face and Gesture Recognition, Nara, Japan, 2004.130-135.
[16] Milan Sonka. Vaclav Hlavac. Roger Boyle. Image Processing, Analysis and Machine Vision(Second Edition).USA:Brooks/Cole.
[17] 钟志光，卢勇．Visual C++.net数字图像处理与解析．北京：清华大学山版社，2003．
[18] 易丛琴，李丽，孟传良．基于肤色分割的彩色图像的人脸检测．西安交通大学学报，2003，32(4)：65～68．
[19] 章毓晋．图像处理和分析基础．北京：高等教育出版社，2002．
[20] 吴东晖．智能移动机器人的视觉导航技术：(博士论文)．杭州：浙江大学，2006．
[21] YaoquanYang, Jie Bai, Ruili Tian, et al. A Vehicle License Plate Recognition System based on fixed color collocation. Proc. IEEE the Fourth International Conference on Machine Learning and Cybernetics. Guangzhou. 2005: 18-21.
[22] 龚声蓉等．数字图像处理与分析．北京：清华大学出版社，2006．
[23] Hu Yuchen, Chang Chinchen. Edge Detection Using Block Truncation Coding. International Journal Pattern Recognition and Artificial Intelligence, 2003,17(6): 951-966.
[24] Zhang Y J.A survey on evaluation methods for image segmentation. Pattern Recognition, 2000.29 (8):1335-1346
[25] Chalana V, Kim Y. A Methodology for Evaluation of Boundary Detection on Medical Images. IEEE Trans on Medical Imaging, 1999,16(5):642-652.
[26] 欧珊瑚，王倩丽．Visual C++.NET数字图像处理技术与应用．北京：清华大学出版社，2004．
[27] 栾新，朱铁一．不规则类圆形目标图象识别新策略．中国图象图形学报，1999，4(3)，A：202-206．
[28] 王听，曲波．一种不规则封闭形状的轮廓跟踪算法．南京航空航天大学报，2003，28(4)：58—60．
[29] 韩玉杰等．计算机控制激光扫描识别木段横截面图形的研究．木材工业，1999，(1)：13—14．
[30] 邵峰晶，于忠清．数据挖掘——原理与算法．北京：中国水利水电出版社，2003．
[31] 李国辉，武德峰，付畅俭．多媒体知识及其获取．第十三届全国多媒体技术学术会议论文集，宁波，2004：451～456．
[32] 杨枝灵，王开等．Visual C++数字图像获取处理及实践应用．北京：人民邮电出版社，2003．7382504-573．
[33] 崔屹．图象处理与分析——数学形态学方法及应用．北京：科学出版社，2000．
[34] 刘立波．数学形态学在二值图象处理与分析中的研究及应用：(硕士学位论文)．西安：西安交通大学．2002．
[35] Bruce Eckel. Thinking in C++(Second Edition)Volume One Introduction to Standard C++.MindView, Inc, 2000.
[36] TehChuan Chen, KuoLiang Chung. An Efficient Randomized Algorithm for Detecting Circles Computer Vision and Image Understanding, 2001,83(2).172-191.
[37] Rosenfield A. Computer Vision-Past Present and Future. Information Science, 1999:57-58: 291-243.
[38] 童雀菊等．用图像处理法采集原木形状参数的研究．林业科学，2005，34(3)：87—95．
[39] 冯莉．原木材积计算机视觉检测系统的研究：(硕士学位论文)．哈尔滨：尔北林业大学2002(3)．