城市生态修复景观布局特征智能提取方法仿真

Urban Ecological Restoration Landscape Layout Feature Intelligent Extraction Method Simulation

为解决当前城市生态修复景观布局特征提取方法中存在的特征提取准确性差、耗时长的问题,提出基于多指标的城市生态修复景观布局特征智能提取方法。采用遥感数据采集的方式采集研究城市生态修复景观布局图像,并对图像进行预处理,利用处理结果构建一个加权无向图。通过蚁群法实现多层图的划分,判断图划分的归一化割值是否停止变化,如果结果为是,则将获取的最佳图划分映射至原图像,得到最终的图像分割结果。基于城市景观布局图像分割结果,分别将城市生态修复中的景观总面积、景观总周长、景观的分维数、景观的破碎化指数和嵌块体内缘比提取出来。仿真结果表明,所提方法特征提取耗时平均为0.24h,所得特征值与实际情况拟合程度高于当前方法。

In this paper,a method to intelligently extract landscape layout features in urban ecological restoration based on multi-criteria was presented. The remote sensing data collection method was used to collect and research the landscape layout image in urban ecological restoration,and then the image was preprocessed. Meanwhile,the processing result was used to construct a weighted undirected graph. Moreover,the ant colony method was used to achieve the division of the multi-level graph and thus to determine whether the change of normalized cut value of graph stopped. If the result is true,then the best graph division was mapped to the original image to obtain the final result of image segmentation. Based on the segmentation result of urban landscape layout image,the total area of landscape,the total perimeter of landscape,the fractal dimension of landscape,the fragmentation index of landscape and the ratio for interior edge of abaculus were extracted from urban ecological restoration. Form simulation results,we can see that the average time consumption of feature extraction is 0.24 h. Meanwhile,the fitting degree between the obtained feature value and the actual situation is higher than that of current method.