摘要
在工程施工中,经常需要测量施工现场的距离参数,利用可视化工程管控系统获取图像,继而采用双目测距算法对图像进行处理是一种有效解决途径。在双目测距算法中,对图像直接进行视差估计,其精度有时难以满足要求。为解决这一问题,对基于POCS图像信息增强算法进行了改进,提出了基于局部信息增强的双目测距算法,实现了信息增强。通过融入差异值算法,信息增强后的图像细节信息更为丰富,提高了视差估计的精度。由于只对图像的局部信息进行增强,计算机运算量少,耗时更短。实验表明,随着增强力度的提高,双目测距的误差逐渐降低,在分辨率有限条件下,提高了测量精度。同时,突破了定焦摄像机测量距离短的限制。
In engineering construction,it is often necessary to measure the distance parameters of the construction site.Using the visual management and contral system to acquire images,and then employing the binocular ranging algorithm to process the image is an effective solution.In the binocular ranging algorithm,the image is directly subjected to disparity estimation,and the accuracy requirement is sometimes difficult to be satisfied.In order to solve this problem,the POCS image information enhancement algorithm is improved,and the binocular ranging algorithm based on local information enhancement is proposed to realize information enhancement.By incorporating the difference value algorithm,the image details of the information enhancement are more abundant,and the accuracy of the disparity estimation is improved.Since only the local information of the image is enhanced,the computer has less computation and takes less time.Experiments show that with the increase of the enhancement intensity,the error of binocular ranging is gradually reduced,and the measurement accuracy is improved under the condition of limited resolution.At the same time,the limit of the short distance measurement of the fixed focus camera was broken.
作者
何涛
邱捷
张维
He Tao;Qiu Jie;Zhang Wei(Shenzhen Nanshan Power Supply Bureau,Shenzhen 518000,China)
出处
《电子测量技术》
2020年第9期88-93,共6页
Electronic Measurement Technology
基金
深圳供电局有限公司科技项目(90000KK5218035)资助
关键词
双目视觉
局部信息增强
视差估计
差异值
可视化
测距
binocular vision
local information enhancement
disparity estimation
differential value
visualization
ranging
