城市街道景观高差异特征智能识别方法仿真

Simulation of Intelligent Recognition Method for Urban Street Landscape High Difference Characteristics

针对当前方法在对城市街道景观高差异特征进行识别时存在识别准确率较低、错误率较高、耗时较长的缺点,提出一种基于增强算子的城市街道景观高差异特征智能识别方法,通过对采集获得的原始城市街道景观灰度图像进行包括膨胀运算、腐蚀运算、开闭运算以及高低帽运算等一系列形态学滤波,实现对城市街道景观灰度图像的预处理。在对城市街道景观图像进行预处理基础上,将城市街道景观图像划分为若干个等面积子区域,并计算划分后各个子区域的像素灰度值以及节点景观图像像素点分布密度平均值,通过设置综合均值阈值对像素点是否处于城市街道景观图像边缘进行判断,依据判断结果对城市街道景观高差异特征进行智能识别。仿真结果表明,所提方法能够实现城市街道景观高差异特征的高效、准确识别。

Due to low recognition accuracy and high error rate in identifying the significant difference characteristics in urban street landscape, this article proposes an intelligent recognition method for the significant difference characteristics of urban street landscape based on enhanced operator. After getting original grayscale image of urban street landscape, we carried out a series of morphological filtering including expansion operation, corrosion operation, opening and closing operation, and Top-hat and Bottom-hat operation, so that the preprocessing of gray image in urban street landscape was realized. On this basis, we divided the urban street landscape image into several equal-area sub-regions. Meanwhile, we calculated the pixel gray value of each sub-region and the mean value of pixel distribution density of node landscape image. Moreover, we judged whether the pixel was at the edge of urban street landscape image by setting the comprehensive mean threshold. Finally, we intelligently recognized the significant difference characteristics of urban street landscape according to judgment result. Simulation results show that the proposed method can achieve high-efficient and accurate recognition for significant difference characteristics of urban street landscape.