摘要
针对经纬仪测量图像评估的特殊性,根据序列测量图像目标识别一般采用模板匹配的原理,提出了一种基于相关系数的测量图像评估方法.该方法采用对背景区域和目标区域的相关系数分别评估的方式,建立了主次波峰比评估指标,并引入了平均梯度和散度指标.实验证明:本方法能够真实地反映测量图像质量对目标识别算法的适应能力,其评估结果准确、可靠.
For the particularity of the evaluation of the image quality measured by an optical theodolite, a new evaluation method is proposed based on correlation coefficients, which uses the fact that image sequence often uses template matching. The method calculates the correlation coefficients of the background and target regions, constructs the evaluation parameters using the primary and secondary peak ratios, and introduces the average grad and divergence parameters. Experimental result shows that the method has the ability to adapt the quality variety of measured images, and the evaluated results are correct and credible.
出处
《光电技术应用》
2007年第5期81-84,共4页
Electro-Optic Technology Application
关键词
测量图像
评估
模板匹配
目标
背景
measured image
evaluation
template matching
target
background
