顾及特征离散程度的SEaTH特征优化选择方法

doi:10.11947/j.AGCS.2024.20220571

摘要/Abstract

摘要： 特征选择是面向对象信息提取的关键步骤之一。本文针对分离阈值(separability and threshold,SEaTH)这一特征选择方法在实际应用中存在的不足,例如未考虑特征值的离散程度,仅利用J-M距离评判单一特征,特征间可能存在较强相关性,以及无法有效确定出分类顺序,提出了一种改进的SEaTH算法(optimized SEaTH,OPSEaTH)。OPSEaTH算法首先在J-M距离基础上构建了一类特征评价指标(E值),有效解决了特征值的离散度问题;然后,基于E值构建出特征组合评价指标(C_e值),可有效评估得到每种地物的最佳特征组合并自动确定出地物的分类顺序;最后基于eCognition等分类器可完成对地物对象的最终有效分类。利用高分二号遥感影像数据对本文方法进行了测试,并将结果分别与SEaTH算法、DPC、OIF和最近邻分类器的分类结果进行了对比,结果表明:OPSEaTH算法不仅能有效降低特征维数、优化特征空间,还能够对分类顺序进行自动化合理确定,总体精度和Kappa系数及其他精度指标,均显著优于基于SEaTH算法的特征选择结果。本文方法无论从特征降维效果、分类结果精度还是计算效率方面均优于DPC、OIF和最近邻分类器结果。OPSEaTH是一种更优的特征选择方法。

关键词: SEaTH算法, 特征选择, 离散系数, 特征组合, 分类顺序, 改进SEaTH算法

Abstract: Feature selection is one of the key steps in object-oriented information extraction. In view of the fact that the separability and threshold (SEaTH), a feature selection method, does not consider the discrete degree of eigenvalues, only uses the J-M distance to judge a single feature, there may be strong differences between features, and the inability to effectively determine the limitations of the classification order in practical application. Therefore, the extraction of ground objects cannot achieve the optimal effect, the extraction rules of ground objects are also complex, and the portability of the classification model is still poor. To solve these problems, an improved SEaTH algorithm (optimized SEaTH, OPSEaTH) was developed in this study. First, a feature evaluation index (E) is constructed by OPSEaTH based on J-M distance, which can effectively solve the dispersion of eigenvalues. Further, a feature combination evaluation index (C_e) is constructed based on E value, which can effectively evaluate the best feature combination of each feature and automatically determine the classification order of features. Then, the effective classification of feature objects can be completed based on eCognition and other classifiers. In this study, the new algorithm is tested by using GF-2 remote sensing image data, and compared with the classification results of SEaTH algorithm, DPC (density peaks cluster), OIF (optimal index factor), and the nearest neighbor classifier, respectively. The results show that: OPSEaTH algorithm can not only effectively reduce the feature dimension and optimize the feature space, but also automatically and reasonably determine the classification order. The overall accuracy, Kappa coefficient and other accuracy indexes of the OPSEaTH algorithm are significantly better than the feature selection results based on SEaTH algorithm. In addition, the OPSEaTH algorithm is superior to DPC, OIF and the nearest neighbor classifier in terms of feature dimension reduction effect, classification accuracy and computational efficiency. OPSEaTH algorithm is a better feature selection method.

Key words: SEaTH algorithm, feature selection, dispersion coefficient, feature combination, classification order, optimized SEaTH algorithm

中图分类号:

P228

瞿伟, 王宇豪, 王乐, 李久元, 李达. 顾及特征离散程度的SEaTH特征优化选择方法[J]. 测绘学报, 2024, 53(1): 20-35.

QU Wei, WANG Yuhao, WANG Le, LI Jiuyuan, LI Da. An feature optimization selection method of SEaTH considering discretization degree[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(1): 20-35.

参考文献

[1] 洪亮, 冯亚飞, 彭双云, 等. 面向对象的多尺度加权联合稀疏表示的高空间分辨率遥感影像分类[J]. 测绘学报, 2022, 51(2):224-237. DOI: 10.11947/j.AGCS.2022.20190290. HONG Liang, FENG Yafei, PENG Shuangyun, et al. Classification of high spatial resolution remote sensing imagery based on object-oriented multi-scale weighted sparse representation[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(2):224-237. DOI: 10.11947/j.AGCS.2022.20190290.
[2] XU Kunpeng, ZHAO Zhao, LI Kun, et al. Estimation of crop biomass using GF-3 polarization SAR data based on genetic algorithm feature selection[J]. Journal of Geodesy and Geoinformation Science, 2020,3(4): 126-136.
[3] ZHENG Tong, LEI Peng, WANG Jun. A hybrid features based detection method for inshore ship targets in SAR imagery[J]. Journal of Geodesy and Geoinformation Science, 2023, 6(1):95-107.
[4] TANG J, ALELYANI S, LIU H. Feature selection for classification: a review[J]. Data Classification: Algorithms and Applications, 2014: 37.
[5] LI Wei, LU Chan, CAO Shiyi. Dynamic remote sensing monitoring of land cover in ecological protection area based on GIS technology[J]. International Journal of Environmental Technology and Management, 2021, 24(1-2): 77-92.
[6] ZHAO J, ZHOU Q, CHEN Y, et al. Fusion of visible and infrared images using saliency analysis and detail preserving based image decomposition[J]. Infrared Physics & Technology, 2013, 56: 93-99.
[7] REN Zhiwei, WU Lingda. Hyperspectral intrinsic image decomposition based on automatic subspace partitioning[J]. Laser & Optoelectronics Progress, 2018, 55(10): 103004.
[8] 苏红军. 高光谱遥感影像降维: 进展、挑战与展望[J]. 遥感学报, 2022, 26(8): 1504-1529. SU Hongjun. Dimensionality reduction for hyperspectral remote sensing: advances, challenges, and prospects[J]. National Remote Sensing Bulletin, 2022, 26(8): 1504-1529.
[9] JANG S W, LEE Sanghong. Feature selection based on Euclid distance and neuro-fuzzy system[J]. Journal of Advances in Information Technology, 2020: 155-160.
[10] JIA Sen, TANG Guihua, ZHU Jiasong, et al. A novel ranking-based clustering approach for hyperspectral band selection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(1): 88-102.
[11] HUANG W, ZHANG Y, YU Y, et al. Historical data-driven risk assessment of railway dangerous goods transportation system: comparisons between entropy weight method and scatter degree method[J]. Reliability Engineering & System Safety, 2021, 205: 107236.
[12] GAO Yan, MARPU P, NIEMEYER I, et al. Object-based classification with features extracted by a semi-automatic feature extraction algorithm-SEaTH[J]. Geocarto International, 2011, 26(3): 211-226.
[13] 余晓敏, 湛飞并, 廖明生, 等. 利用改进SEaTH算法的面向对象分类特征选择方法[J]. 武汉大学学报(信息科学版), 2012, 37(8): 921-924. YU Xiaomin, ZHAN Feibing, LIAO Mingsheng, et al. Object-oriented feature selection algorithms based on improved SEaTH algorithms[J]. Geomatics and Information Science of Wuhan University, 2012, 37(8): 921-924.
[14] 张群, 赵超英. 基于面向对象的高分遥感数据甘肃黑方台黄土滑坡半自动识别[J]. 灾害学, 2017, 32(3):210-215. ZHANG Qun, ZHAO Chaoying. Semiautomatic object-oriented loose landslide recognition based on high resolution remote sensing images in Heifangtai, Gansu[J]. Journal of Catastrophology, 2017, 32(3):210-215.
[15] 王春燕, 王刘明, 张媛, 等. 基于SEaTH的决策树方法在区域尺度土地覆被分类中的应用[J]. 兰州大学学报(自然科学版), 2019, 55(1): 11-18. WANG Chunyan, WANG Liuming, ZHANG Yuan, et al. Application of decision tree method based on SEaTH in land cover classification at regional scale[J]. Journal of Lanzhou University (Natural Sciences), 2019, 55(1): 11-18.
[16] 王贺, 陈劲松, 余晓敏. 面向对象分类特征优化选取方法及其应用[J]. 遥感学报, 2013, 17(4):816-829. WANG He, CHEN Jinsong, YU Xiaomin. Feature selection and its application in object-oriented classification[J]. Journal of Remote Sensing, 2013, 17(4):816-829.
[17] 何志文, 李夕海, 刘代志, 等. 基于相关性分析的特征选择方法研究[J]. 核电子学与探测技术, 2005, 25(6):729-732, 749. HE Zhiwen, LI Xihai, LIU Daizhi, et al. A study on methods of feature selection based on the correlation analysis[J]. Nuclear Electronics & Detection Technology, 2005, 25(6):729-732, 749.
[18] SCHOBER P, BOER C, SCHWARTE L A. Correlation coefficients: appropriate use and interpretation[J]. Anesthesia & Analgesia, 2018, 126(5): 1763-1768.
[19] LEE RODGERS J, ALAN NICEWANDER W. Thirteen ways to look at the correlation coefficient[J]. The American Statistician, 1988, 42(1): 59-66.
[20] MARPU P R, NIEMEYER I, NUSSBAUM S, et al. A procedure for automatic object-based classification[M]//Lecture Notes in Geoinformation and Cartography. Berlin: Springer, 2008: 169-184.
[21] FAN Jinlong, DEFOURNY Pierre, DONG Qinghan, et al. Sent2Agri system based crop type mapping in Yellow River irrigation area[J].Journal of Geodesy and Geoinformation Science,2020,3(4):110-117.
[22] AO Y, WANG J, ZHOU M, et al. Fully convolutional networks for street furniture identification in panorama images[J]. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2019, XLII-2/W13: 13-20.
[23] PAL N R, PAL S K. A review on image segmentation techniques[J]. Pattern recognition, 1993, 26(9): 1277-1294.
[24] BENZ U C, HOFMANN P, WILLHAUCK G, et al. Multi-resolution, object-oriented fuzzy analysis of remote sensing data for GIS-ready information[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2004, 58(3-4): 239-258.
[25] ZHAO Yi, JIANG Mi, MA Zhangfeng. Integration of SAR polarimetric features and multi-spectral data for object-based land cover classification[J]. Journal of Geodesy and Geoinformation Science, 2019, 2(4):64-72.
[26] 魏立飞, 余铭, 钟燕飞, 等. 空-谱融合的条件随机场高光谱影像分类方法[J]. 测绘学报, 2020, 49(3):343-354. DOI: 10.11947/j.AGCS.2020.20190042. WEI Lifei, YU Ming, ZHONG Yanfei, et al. Hyperspectral image classification method based on space-spectral fusion conditional random field[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(3):343-354. DOI: 10.11947/j.AGCS.2020.20190042.
[27] 王之, 刘超, 刘秀菊, 等. 基于SEaTH算法的芦山地震无人机低空遥感影像信息对象级分类[J]. 地震研究, 2018, 41(2):173-179. WANG Zhi, LIU Chao, LIU Xiuju, et al. Study on the object-based classification of low-altitude UAV remote sensing image of the Lushan earthquake based on the SEaTH algorithm[J]. Journal of Seismological Research, 2018, 41(2):173-179.
[28] 张勤, 赵超英, 陈雪蓉. 多源遥感地质灾害早期识别技术进展与发展趋势[J]. 测绘学报, 2022, 51(6):885-896. DOI: 10.11947/j.AGCS.2022.20220132. ZHANG Qin, ZHAO Chaoying, CHEN Xuerong. Technical progress and development trend of geological hazards early identification with multi-source remote sensing[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(6):885-896. DOI: 10.11947/j.AGCS.2022.20220132.