基于词袋特征的空心村高分影像建筑物解译模型

发布时间：2018-04-05 18:45

  本文选题：空心村　切入点：词袋模型　出处：《农业机械学报》2017年06期

【摘要】：如何利用高分影像构建自动解译模型是快速高效获取空心村建筑物的关键,对空心村调查研究具有重要意义。针对传统目视解译需要专业知识,效率低、工作量大的问题,提出一种基于词袋特征的空心村高分影像建筑物解译模型。首先,对比了多种影像特征提取方法;然后,选取词袋特征(Bo W)和支持向量机(SVM)构建建筑物自动解译模型;最后,为检验方法的有效性,选取空心村高分影像构建了建筑物样本库,并基于该样本库进行实验研究。结果表明本文方法的分类准确度可以达到0.86,所提方法可用于空心村内建筑物自动解译,具有较高的实用价值。
[Abstract]:How to construct an automatic interpretation model by using high score image is the key to acquire hollow village buildings quickly and efficiently, which is of great significance to the investigation and research of hollow villages.In order to solve the problem that traditional visual interpretation requires professional knowledge, low efficiency and heavy workload, a high score image interpretation model of hollow village based on word bag feature is proposed.Firstly, several image feature extraction methods are compared. Then, the automatic interpretation model of building is constructed by selecting word bag feature (Bo W) and support vector machine (SVM). Finally, in order to verify the validity of the method,The building sample database is constructed by selecting the hollow village high score image, and the experimental research is carried out based on the sample library.The results show that the classification accuracy of this method can reach 0.86. The proposed method can be applied to the automatic interpretation of buildings in hollow villages and has high practical value.
【作者单位】： 西南交通大学地球科学与环境工程学院;四川省土地统征整理事务中心;四川大学水力学与山区河流开发保护国家重点实验室;四川大学水利水电学院;
【基金】：“十二五”国家科技支撑计划项目(2014BAL01B04)
【分类号】：P237;TP751
，

本文编号：1715994