第13卷第2期 智能系统学报 Vol.13 No.2 2018年4月 CAAI Transactions on Intelligent Systems Apr.2018 D0:10.11992/tis.201609012 网络出版地址：http:/kns.cnki.net/cms/detail/23.1538.TP.20170626.1740.016.html 直线截距直方图城区遥感图像多阈值分割 吴诗婚，吴一全25，周建江 (1.南京航空航天大学电子信息工程学院，江苏南京211106,2.城市空间信息工程北京市重点实验室，北京100038： 3.江西省数宇国土重点实验室，江西南昌330013,4.江苏省大数据分析技术重点实验室，江苏南京210044：5.浙江 省信号处理重点实验室，浙江杭州310023) 摘要：阈值分割简单有效，但现有的单阈值方法对城区图像分割效果不佳，难以取得令人满意的结果。为了快速准 确地对城区遥感图像进行分割，本文提出了基于直线截距直方图倒数灰度嫡和人工蜂群优化(artificial bee colony op- timization,.ABC)的多阈值分割方法。首先，给出直线截距直方图的定义并建立城区遥感图像的直线截距直方图：然 后，计算该直方图倒数灰度嫡的大小，推导出其单阈值选取公式：最后，将其推广到多阈值选取，并利用人工蜂群优化 算法，对多个阈值进行快速精确地寻优，以此最终实现城区遥感图像的多阈值分割。实验结果表明，该方法所分割的 图像中多目标的形状、边缘更为准确，纹理及细节特征更加清晰，且所需运行时间仅为同类多阈值分割方法的25% 是一种行之有效的城区遥感图像分割方法。 关键词：城区提取；遥感图像；图像分割：阈值化：多阈值选取；直线截距直方图；倒数灰度嫡；人工蜂群优化 中图分类号：TP751.1:P237文献标志码：A文章编号：1673-4785(2018)02-0227-09 中文引用格式：吴诗婳，吴一全，周建江.直线截距直方图城区遥感图像多阈值分割.智能系统学报，2018.13(2)：227-235 英文引用格式：WU Shihua,WU Yiquan,ZHOU Jianjiang.Multi--level thresholding for remote sensing image of urban area based on line intercept histogramJl.CAAI transactions on intelligent systems,2018,13(2):227-235. Multi-level thresholding for remote sensing image of urban area based on line intercept histogram WU Shihua',WU Yiquan245,ZHOU Jianjiang' (1.College of Electronic and Information Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China; 2.Beijing Key Laboratory of Urban Spatial Information Engineering,Beijing 100038,China;3.Jiangxi Province Key Laboratory for Digital Land,Nanchang 330013,China;4.Jiangsu Key Laboratory of Big Data Analysis Technology,Nanjing 210044,China;5.Zhejiang Province Key Laboratory for Signal Processing,Hangzhou 310023,China) Abstract:Threshold segmentation is a kind of simple and effective method,however,the existing single-threshold method is hard to realize satisfactory effect in segmenting the images of urban area.In order to segment the remote sens- ing images of urban area quickly and accurately,a multi-threshold segmentation method based on straight-line intercept histogram,reciprocal grayscale entropy and Artificial Bee Colony(ABC)Optimization was proposed in the paper. Firstly,the straight-line intercept histogram was defined and the straight-line intercept histogram of the urban remote sensing image was established;then the value of the reciprocal grayscale entropy of the histogram was calculated and the single-threshold selection formula was deduced;finally,the application was popularized to multi-threshold selection, ABC Optimization algorithm was utilized for precise optimization of many thresholds,so as to finally realize the multi- threshold segmentation of urban remote sensing images.A large number of experiments show that,the multi-object shape and edge in the images segmented by the method are more accurate,the textures and details are more explicit,in addition,its running time is only 25%of other similar multi-threshold segmentation methods.This is a kind of effective method for segmenting the remote sensing images of urban area. Keywords:extraction of urban area;remote sensing image;image segmentation;thresholding,multi-level threshold se- lection;straight-line intercept histogram;reciprocal grayscale entropy;optimization of artificial bee colony 收稿日期：2016-09-28.网络出版日期：2017-06-26. 随着遥感技术的飞速发展，利用卫星遥感和飞 基金项目：国家自然科学基金项目(61573183)：城市空间信息工程 机遥感等方式实时获取的地物图像质量也越来越 北京市重点实验室开放基金项目(2014203)：江西省数 字国土重点实验室开放基金项目(DLLJ201412):江苏 高。从人工地物的遥感图像中提取信息，可以避免 省大数据分析技术重点实验室开放基金项目 (KXK1403):浙江省信号处理重点实验室开放基金项目 传统的实地勘测，大大提高工作效率。城区作为遥 (ZJKL6SP.OP2014-02):江苏高校优势学科建设工程 感图像中一类重要的人工地物目标，其自动提取在 项目(2012). 通信作者：吴一全.E-mail:nuaaimage@163.com. 城市规划、地理信息系统更新、数字化城市以及军DOI: 10.11992/tis.201609012 网络出版地址: http://kns.cnki.net/kcms/detail/23.1538.TP.20170626.1740.016.html 直线截距直方图城区遥感图像多阈值分割 吴诗婳1 ，吴一全1,2,3,4,5，周建江1 （1. 南京航空航天大学 电子信息工程学院，江苏 南京 211106; 2. 城市空间信息工程北京市重点实验室，北京 100038; 3. 江西省数字国土重点实验室，江西 南昌 330013; 4. 江苏省大数据分析技术重点实验室, 江苏 南京 210044; 5. 浙江 省信号处理重点实验室, 浙江 杭州 310023） 摘 要：阈值分割简单有效，但现有的单阈值方法对城区图像分割效果不佳，难以取得令人满意的结果。为了快速准 确地对城区遥感图像进行分割，本文提出了基于直线截距直方图倒数灰度熵和人工蜂群优化 (artificial bee colony op￾timization, ABC) 的多阈值分割方法。首先，给出直线截距直方图的定义并建立城区遥感图像的直线截距直方图；然 后，计算该直方图倒数灰度熵的大小，推导出其单阈值选取公式；最后，将其推广到多阈值选取，并利用人工蜂群优化 算法，对多个阈值进行快速精确地寻优，以此最终实现城区遥感图像的多阈值分割。实验结果表明，该方法所分割的 图像中多目标的形状、边缘更为准确，纹理及细节特征更加清晰，且所需运行时间仅为同类多阈值分割方法的 25%， 是一种行之有效的城区遥感图像分割方法。 关键词：城区提取；遥感图像；图像分割；阈值化；多阈值选取；直线截距直方图；倒数灰度熵；人工蜂群优化 中图分类号：TP751.1；P237 文献标志码：A 文章编号：1673−4785(2018)02−0227−09 中文引用格式：吴诗婳, 吴一全, 周建江. 直线截距直方图城区遥感图像多阈值分割[J]. 智能系统学报, 2018, 13(2): 227–235. 英文引用格式：WU Shihua, WU Yiquan, ZHOU Jianjiang. Multi-level thresholding for remote sensing image of urban area based on line intercept histogram[J]. CAAI transactions on intelligent systems, 2018, 13(2): 227–235. Multi-level thresholding for remote sensing image of urban area based on line intercept histogram WU Shihua1 ，WU Yiquan1,2,3,4,5 ，ZHOU Jianjiang1 (1. College of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China; 2. Beijing Key Laboratory of Urban Spatial Information Engineering, Beijing 100038, China; 3. Jiangxi Province Key Laboratory for Digital Land, Nanchang 330013, China; 4. Jiangsu Key Laboratory of Big Data Analysis Technology, Nanjing 210044, China; 5. Zhejiang Province Key Laboratory for Signal Processing, Hangzhou 310023, China) Abstract: Threshold segmentation is a kind of simple and effective method, however, the existing single-threshold method is hard to realize satisfactory effect in segmenting the images of urban area. In order to segment the remote sens￾ing images of urban area quickly and accurately, a multi-threshold segmentation method based on straight-line intercept histogram, reciprocal grayscale entropy and Artificial Bee Colony (ABC) Optimization was proposed in the paper. Firstly, the straight-line intercept histogram was defined and the straight-line intercept histogram of the urban remote sensing image was established; then the value of the reciprocal grayscale entropy of the histogram was calculated and the single-threshold selection formula was deduced; finally, the application was popularized to multi-threshold selection, ABC Optimization algorithm was utilized for precise optimization of many thresholds, so as to finally realize the multi￾threshold segmentation of urban remote sensing images. A large number of experiments show that, the multi-object shape and edge in the images segmented by the method are more accurate, the textures and details are more explicit, in addition, its running time is only 25% of other similar multi-threshold segmentation methods. This is a kind of effective method for segmenting the remote sensing images of urban area. Keywords: extraction of urban area; remote sensing image; image segmentation; thresholding; multi-level threshold se￾lection; straight-line intercept histogram; reciprocal grayscale entropy; optimization of artificial bee colony 随着遥感技术的飞速发展，利用卫星遥感和飞 机遥感等方式实时获取的地物图像质量也越来越 高。从人工地物的遥感图像中提取信息，可以避免 传统的实地勘测，大大提高工作效率。城区作为遥 感图像中一类重要的人工地物目标，其自动提取在 城市规划、地理信息系统更新、数字化城市以及军 收稿日期：2016−09−28. 网络出版日期：2017−06−26. 基金项目：国家自然科学基金项目 (61573183)；城市空间信息工程 北京市重点实验室开放基金项目 (2014203)；江西省数 字国土重点实验室开放基金项目 (DLLJ201412)；江苏 省大数据分析技术重点实验室开放基金项目 (KXK1403)；浙江省信号处理重点实验室开放基金项目 (ZJKL_6_SP-OP2014-02)；江苏高校优势学科建设工程 项目 (2012). 通信作者：吴一全. E-mail：nuaaimage@163.com. 第 13 卷第 2 期 智 能 系 统 学 报 Vol.13 No.2 2018 年 4 月 CAAI Transactions on Intelligent Systems Apr. 2018