为解决结构光三维扫描中线式激光中心线的提取方法多义性,分析激光线的特点,综合考虑激光中心线精度、算法提取速度及图像噪音3个因素,提出一种线式激光中心的二次通道提取算法。以激光线在图像中服从高斯分布为基础,对图像的高斯分布峰值进行两次概率估算,不断对激光中心坐标进行逼近,滤去其背景图案,获得亚像素级激光中心带,建立基于高斯分布的图像通道提取算法。将仿真结果与其它算法进行比较,比较结果表明,该算法能快速收敛并准确得出激光中心坐标域,优于其它算法。
To solve the polysemy of extraction method of linear laser center line from the structural,optical and three-dimensional(3D)scan,the characteristics of the laser line were analyzed,meanwhile,considering the three factors synthetically,including the accuracy of the laser center line,the speed of extraction algorithm,image and noise,the extraction algorithm of the secondary channel of linear laser center was then proposed.The fact that the Gauss distribution was submitted by the image of linear laser was taken as the foundation of this algorithm,the Gauss distribution peak of the image was estimated on the basis of probability twice.The laser center coordinates were evacuated ceaselessly and the background image was filtered.The sub-pixel center with laser was acquired and the extraction algorithm was established based on the Gauss distribution channel.Compared the simulation results with that of the other algorithms,the algorithm has quicker convergence and more accurate laser center coordinates.