一种基于Harris角点检测的SVG转换可视化效果评价方法

A Harris Corner Detection Based Evaluation Method for SVG Visualization Conversion Effect

SVG作为移动互联时代获得各种丰富图形信息的有效表达方法,已成为LBS中空间信息表达与描述的热点应用之一,但目前的研究大多没有考虑多源数据格式向SVG格式转换后的图像质量评价问题。针对多源空间数据向SVG格式转换方法和转换效果评价技术,本文设计了栅格数据和矢量数据向SVG格式转换的表达方法,分别利用直接载入栅格图像和编程实现矢量数据转换的方法实现了多源空间数据向SVG格式的有效转换;从定性与定量相结合的角度分析了转换前后的图像质量,设计了角点检测评价指标和流程,提出了基于Harris角点检测的SVG转换前后图形几何质量的评价方法。实验证明:SVG格式比较适用于位置服务中对空间数据尤其是矢量数据的表达,经过SVG转换后的图像,无论是栅格数据还是矢量数据都具有较高的图像质量,且栅格数据的转换效果相对略差于矢量数据;验证了一种基于SVG的移动位置空间服务信息表达和组织模式,对普及和强化SVG在位置服务系统中的应用具有一定的推动作用。

With the development of mobile communication technologies, the demands of location-based spatial information services are more and more extensive in order for scientists to effectively obtain rich spatial informa- tion in mobile internet era anytime and anywhere, and SVG （Scalable Vector graphics） is becoming a hot applica- tion on data representation and organization in Location Based Services （LBS）. However, few past researches fo- cused on image quality assessment of the conversion from different spatial data format to SVG format. So in this context, according to the unique demand on SVG_based representation and organization for spatial data in LBS, this research put forward the conversion method and quality assessment method from multiple formats to SVG format. Firstly, by using of the directly loaded raster images and by running the vector data conversion pro- gramme, a valid conversion is achieved from multi-source spatial data formats to SVG format. Secondly, from the combination of qualitative and quantitative view, the comparison between the image quality before and after the conversion is analyzed. And, based on the design of comer detection index and evaluation processes, a Har- ris comer detector method is proposed and verified to evaluate the transform result. The test shows that SVG_based representation is suited for spatial data in LBS, especially for the vector data, and the SVG image af- ter conversion for either raster or vector data has a high image quality. As the JPG raster image is concerned, compared with the original image, the accuracy of the converted SVG image changed little, but the resolution of the converted SVG image is less than the original one. On the other hand, the accuracy and clarity of the convert- ed vector data is similar to the original vector data. And the conversion of raster data is less effective than that of vector data, which shows that the comer detection technique on SVG transform is feasible. Further, it verifies a reliable SVG based mobile location service information expression and spatial organization model, and helps to popularize and strengthen SVG application in LBS.