摘要
针对当前大多的双目视觉测量算法都是以PC为平台实现的,但这种方式在实时性与便携性方面存在不足。基于嵌入式开发平台,对双目视觉中成熟的Harris角点提取算法与传统的角点匹配算法,结合双目视觉系统,被测物体的特征和嵌入式平台自身的特点,进行算法上的改进与扩展。在不影响整个系统精度的前提下,达到实时,快速的三维测量。最终为解决非接触式的测量方式,提供一个快速的解决方案。
Nowadays many Binocular vision and measurement algorithms are based on PC platform,and however,this pattern has some limitations in real-time capability and portability.Combined with the binocular vision system,the characteristics of the measured object,and the characteristics of embedded system,some improvement or expansion is done on the mature Harris corner detection algorithm and traditional corner matching algorithm,thus to realize fast and real-time measurement of 3D object without any reduction of the system accuracy.Ultimately a quick solution to the non-contact measurement is provided in this paper.
出处
《通信技术》
2011年第10期55-57,60,共4页
Communications Technology
关键词
双目视觉
嵌入式
HARRIS角点
角点匹配
binocular vision
embedded system
Harris corner detection
corner matching
