基于街景图像解译的寒地城市绿视率分析研究--以哈尔滨为例

An Analysis of Green View Index in Cold Region City: A Case Study of Harbin

通过解析街景图像能够自动地测度绿视率,但目前尚无剔除秋冬季街景图像的方法。文章提出使用多来源图像数据,采用SIFT匹配算法判定相似图像的方法剔除秋冬季街景图像。以哈尔滨为例,从整体、区域和街道3个层次,根据POI密度识别城市中心区、郊区以及各功能区,并将绿视率与其叠加分析后得出结论:中心区的绿视率显著高于郊区;单一功能的绿视率排序为服务<居住<工作。绘制不同等级道路的街景绿像热度图,得到不同等级道路绿视率特点,即快速路绿视率最高,其次是主干道和支路,再次是次干路,高速公路最低。对于高等级道路,中央绿化带提升绿视率的作用明显;对于低等级道路,行道树对绿视率的影响较大。

With street view images, researchers can estimate the GVI automatically, However, methods to screen outimages collected in the fall and winter has not been found yet, The paper propose that it is workable to distinguishstreet view images of the fall and winter from multiple suppliers by using SIFT methodology, Taking Harbin as anexample, it makes the analysis at the levels of entire region, specific region and street, The density of POIs is used torecognize the central area and the suburban area, and also used to divide urban spaces of different functions, Byanalysis their GVI features we found that GVI in the central area is higher than that in the suburb and the GVI ranksof different single functions （not mixed） is ordered by working area 〉 living area 〉 service area, We also draw a heatmap by overlying many images, and find the features of different levels of street： the GVI ranks as expressway 〉 mainroad 〉 access road 〉 secondary main road 〉 highway, For higher level roads, the central green belt is the main factorwhich affects the GVI, For lower level roads, roadside trees are the main factor.