基于卷积运算的城市公园绿地聚集度评价

The aggregation degree evaluation of urban park green space based on convolution method

为使公园绿地聚集度计算能够充分反映其辐射效应,提出基于卷积运算的局部网格单元和整个城市公园绿地聚集度评价方法。基于高分辨率遥感影像获取山东省37个主要园林城市公园绿地分布图,利用GIS技术,采用500 m网格将城区网格化;建立3×3绿地聚集度卷积模板,基于卷积运算计算城市网格单元公园绿地聚集度;选择评价因子,依据37个城区公园绿地网格单元聚集度分布现状确定其分级值,建立整个城市公园绿地聚集度评价模型,并对东营市和泰安市进行实例评价。研究结果显示:基于卷积运算的网格单元公园绿地聚集度计算方法能够有效量化相邻网格单元绿地的辐射效应,计算由网格内部及相邻区域绿地共同作用产生的绿地聚集度,其取值范围为0—4;整个城市公园绿地聚集度分为1级(极弱)、2级(弱)、3级(中等)、4级(强)、5级(极强)共5个等级,评价结果与研究区37个城市绿地现状相对应。实例评价结果显示,东营市网格单元公园绿地聚集度主要分布在>0—0.2之间,整个城市公园绿地聚集度为2级;泰安市网格单元公园绿地聚集度以0为主,整个城市公园绿地聚集度为5级。

To fully reflect radiation effects of urban park green space （UPGS） service functions and obtain UPGS aggregation degree （AD） produced by a combination of UPGS in a local area and adjacent ones, the UPGS AD evaluation method for grid units and entire urban area were established based on a convolution operation. Based on high spatial resolution remote sensing images, UPGS maps of 37 cities in Shandong Province were constructed. Then, urban areas were segmented by 500-m grids and GIS technology, thereby obtaining UPGS grid maps. At the same time, a 3x3 AD evaluation convolution mask was constructed and UPGS AD in each grid unit was attained by the convolution operation. Based on the calculation of UPGS AD in each grid unit, two evaluation factors were selected, namely the area ratio of urban space for which AD at each grid was zero, and AD standard deviation of all grids whose AD was larger than zero, for UPGS AD evaluation model construction of the entire urban area. The graded value of the two factors was acquired according to the distribution of UPGS AD in each grid unit of the study area. Using the two factors, the UPGS AD evaluation model of the entire urban area was constructed by a weighted overlay method. Dongying and Tai＇an were taken as typical examples for UPGS AD evaluation. The study result showed that the UPGS AD evaluation method of each grid unit based on convolution calculation could quantify the UPGS radiation effect in adjacent grids well. The method obtained AD values produced by the combination of UPGS in any grid unit and adjacent units. The value of AD ranged from 0 to 4. A value of 0 meant that there was no UPGS in the grid unit and adjacent grids and no AD in the grid unit. With increase of GPGS in the grid unit or adjacent grids, the AD value increased. For AD of 4, all the area in the grid unit and adjacent grids was UPGS, and AD became maximum. The UPGS AD evaluation result of each grid unit could be used to directly guide UPGS planning. The UPGS AD of entire urban area was categorized into five classes, including 1st grade （extremely weak）, 2rid grade （weak）, 3rd grade （ moderate）, dth level （strong） and 5th level （ extremely strong）. The increase of both area ratio of urban space for which AD in each grid was zero and AD standard deviation of all grids whose AD was larger than zero would cause UPGS AD of the entire urban area to rise. However, the level of contrast of UPGS AD corresponded with the 37 cities in the study area, so the result of AD evaluation could be used both for contrast analysis among different cities and to help avoid the phenomenon of going after uniform simply. The evaluation result of two example cities showed that the value of grid-unit PGS AD in Dongying was mainly in the range of 〉0-0.2, and the PGS AD level of the entire urban area was weak. Values of grid-unit PGS AD in Tai＇an were mainly zero, and the PGS AD level of entire urban area was extremely strong.