顾及可信度的OpenStreetMap数据清理

OpenStreetMap(OSM)的数据由大众自发标报,因而存在大量虚假、低质量的数据,数据清理是OSM数据走向实用的重要前提。可信度被多位研究者提出作为OSM数据质量的评价手段,但目前OSM数据清理的研究工作中很少顾及众源地理信息的可信度,因此,本文提出了顾及可信度的OSM数据清理模型。该模型首先计算出顾及用户信誉的OSM数据的可信度,再对数据按照属性进行分类,根据不同属性的数据可信度分布情况设置该属性数据相应的阈值并进行清理,对低于阈值的数据进行人工检查,高于阈值的数据则筛选出用户信誉与可信度不相等以及新建的数据进行人工检查,最后采用OSM真实历史数据中的线对象进行实验,实验结果表明:该模型能够有效地清理出绝大部分的虚假以及低质量数据。

Cleaning Model of OSM Data which Considering the Trustworthiness

To solve the different types of vandalism in OpenStreetMap,this article proposes a cleaning model of OSM data which considering the trustworthiness.Firstly,the model calculates the trustworthiness of OSM data,secondly,classified the data group by attribute,setting the threshold according to the trustworthiness distribution of the corresponding attributes to clean up.Manual inspection of the data which is below the threshold,Selected the data that above the threshold which the user reputation and trustworthiness is not equal as well as the created data to manual inspection,finally,historicalpolyline data from OpenStreetMap is employed for experiments.The results demonstrate that this model can effectively clean out most of the fake and low-quality data.