红外技术, 2017, 39 (9): 824, 网络出版: 2017-12-04
图像聚焦程度对特征提取的影响
Influence of Degree of Image Focusing on Feature Extraction
摘要
纹理性分析是关于图像中局部区域内定量计算像素间关系的。形状是直观描述物体的特征之一,这一特征可以直接用于目标物体的区别中。本文主要以提取红外聚焦图像特征为前提,分析其纹理、面积及周长等特征随聚焦程度变化的趋势。通过实验得出随聚焦程度正相关变化的特征为图像纹理特性中的相关与面积特征,而与之变化负相关的为纹理特征中的二阶矩、熵、对比度。形状特征中的周长与聚焦程度的变化关系较小。在不同的聚焦程度下提取出图像的纹理、形状特征都会有所差异。
Abstract
Texture analysis includes the quantitative calculation of the relationship among pixels in a local area of an image. “Shape” is one of the characteristics of the intuition is tic description of an object, which can be directly used to distinguish the target object. In this paper, the characteristics of “texture,”“area,” and “perimeter” of the infrared focus image are analyzed. The results show positive correlations of image area and image texture with the degree of focusing, while negative correlations exist with regard to the statistical second-order moment, entropy, and contrast of the texture. The relationship between the perimeter of the shape and the degree of focusing is weak. The texture and shape features of the extracted images vary at different degrees of image focus.
郝争辉, 王高, 原东方, 褚星明, 蔚旋, 李咸静. 图像聚焦程度对特征提取的影响[J]. 红外技术, 2017, 39(9): 824. HAO Zhenghui, WANG Gao, YUAN Dongfang, CHU Xngming, WEI Xuan, LI Xianjing. Influence of Degree of Image Focusing on Feature Extraction[J]. Infrared Technology, 2017, 39(9): 824.