论文部分内容阅读
为了提高建筑物提取的自动化程度和精度,提出了一种以分割-分类-优化为主线、利用偏移阴影分析的建筑物全自动提取方法。首先,采用面向对象的多尺度分割方法进行影像初分割;然后,结合支持向量机(SVM)分类,将分割结果分为阴影、植被、建筑物、裸地四大类并提取初始结果;最后,利用相交边界阴影比率准确地验证了建筑物的存在,剔除了无阴影的非建筑物干扰,获取了最终结果。大量的实验结果验证了该方法的有效性,自动化程度得到明显提高。该方法完整度达到85%以上,正确率和综合分数F1均达到90%以上,且仅需要可见光波段影像数据,适用范围广。
In order to improve the degree of automation and the accuracy of building extraction, an automatic building extraction method based on segmentation-classification-optimization and offset shadow analysis is proposed. First of all, the image segmentation is firstly carried out by adopting an object-oriented multi-scale segmentation method. Then, the classification results are divided into four categories: shadows, vegetation, buildings and bare land, and the initial results are extracted according to the support vector machine (SVM) The intersecting boundary shadow ratio is used to verify the existence of the building accurately, and the non-shadow non-building disturbance is eliminated and the final result is obtained. A large number of experimental results verify the effectiveness of the method, the degree of automation has been significantly improved. The method has a completeness of more than 85%, and the correctness rate and the comprehensive score of F1 are all over 90%. Only the visible light band image data is needed, which has wide application range.