针对空间金字塔视觉词袋模型对地形图像分类时忽略颜色信息、纹理表达不明显及特征维度高等问题,提出一种基于空间金字塔模型的DCA特征融合地形分类方法。该方法优化传统空间金字塔模型子区域划分方式,提取地形图像优化后的SPM-BOVW特征、HSV特征、LBP特征;通过DCA算法构建3组变换特征;采用串联将变换特征进行融合。实验结果表明,以融合特征作为支持向量机（SVM）分类器的输入,利用网格参数寻优,最终获得了较高的地形分类精度,说明所提方法在太阳能电站的地形图像分类上具有较好的鲁棒性。
Abstract
Aiming at the problem that the spatial pyramid bag of visual words model does not express clearly the terrain texture of terrain, the color information of the terrain is ignored and high feature dimension, a DCA feature fusion terrain classification algorithm based on the spatial pyramid bag of visual words model is proposed. This method optimizes the sub-region division method of the traditional spatial pyramid model, extracts the optimized SPM-BOVW features, HSV features, and LBP features of the terrain image; constructs three sets of transformation features through the DCA algorithm; and fuses the transformation features in series. The experimental results show that taking the fused features as the input of support vector machines（SVM ）classifier and optimizing through grid parameters, a high accuracy of terrain classification is finally obtained, which shows that the proposed algorithm has good robustness in the terrain classification of solar power station.
关键词
太阳能 /
特征提取 /
支持向量机 /
特征融合 /
地形分类
Key words
solar power /
feature extraction /
support vector machines（SVM） /
feature fusion /
terrain classification
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 李翠明, 王宁, 张晨. 基于改进遗传算法的光伏板清洁分级任务规划[J]. 上海交通大学学报, 2021, 55(9): 1169-1174.
LI C M, WANG N, ZHANG C.Hierarchical mission planning for cleaning photovoltaic panels based on improved genetic algorithm[J]. Journal of Shanghai Jiaotong University, 2021, 55(9): 1169-1174.
[2] COMIN F, SAAJ C.Planetary soil classification based on the analysis of the interaction with deformable terrain of a wheel-legged robot[C]//2015 IEEE/RSJ International Conference on Intelligent Robots & Systems, IEEE. Hamburg, Germany, 2015: 5460-5465.
[3] LAZEBNIK S, SCHMID C, PONCE J.Beyond bags of features: spatial pyramid matching for recognizing natural scene categories[C]//IEEE Society Conference on Computer Vision and Pattern Recognition. New York, USA, 2006: 2169-2178.
[4] 章海兵, 刘士荣, 张波涛, 等. 基于SPM-IBOVW模型的自然场景识别[J]. 上海交通大学学报, 2016, 50(6): 902-909.
ZHANG H B, LIU S R, ZHANG B T, et al.Natural scene recognition based on spatial pyramid integrated bag-of-visual words model[J]. Journal of Shanghai Jiaotong University, 2016, 50(6): 902-909.
[5] CHEN S, LI X, CHI S, et al.Ship target discrimination in SAR images based on BOW model with multiple features and spatial pyramid matching[J]. IEEE access, 2020, 8: 166071-166082.
[6] 毛峡, 李亚豪. 光伏阵列故障检测中的无人机红外图像拼接[J]. 太阳能学报, 2020, 41(3): 262-269.
MAO X, LI Y H.Infrared image stitching of UAV in fault detection of photovoltaic array[J]. Acta energiac solaris sinica, 2020, 41(3): 262-269.
[7] 苗荣慧, 杨华, 武锦龙, 等. 基于图像分块及重构的菠菜重叠叶片与杂草识别[J]. 农业工程学报, 2020, 36(4): 178-184.
MIAO R H, YANG H, WU J L, et al.Weed identification of overlapping spinach leaves based on image sub-block and reconstruction[J]. Transactions of the CSAE, 2020, 36(4): 178-184.
[8] PAN Z, LI Z, FAN H, et al.Feature based local binary pattern for rotation invariant texture classification[J]. Expert systems with applications, 2017, 88: 238-248.
[9] HAGHIGHAT M, ABDEL MOTTALEB M, ALHALABI W.Discriminant correlation analysis: real-time feature level fusion for multimodal biometric recognition[J]. IEEE transactions on information forensics & security, 2016, 11(9): 1984-1996.
[10] 葛芸, 马琳, 储珺. 结合判别相关分析与特征融合的遥感图像检索[J]. 中国图象图形学报, 2020, 25(12): 2665-2676.
GE Y, MA L, CHU J.Remote sensing image retrieval combining discriminant correlation analysis and feature fusion[J]. Journal of image and graphics, 2020, 25(12): 2665-2676.
基金
甘肃省自然科学基金（18JR3RA139）; 国家自然科学基金（51765031）
PDF(1983 KB)
