家具定制系统中基于视觉效果的边缘检测方法

Edge detection method based on visual effects in customized furniture system

针对现有边缘检测算子检测家具图像过程中丢失了原图的视觉效果,而不能应用到家具客户化定制系统中的缺点,提出一种基于新型卷积核的Sobel边缘检测方法.该方法首先将RGB图像按照三原色分割成R图像、G图像、B图像,其次将分割开的图像用文中描述的模板进行离散性加权运算.经实验表明,文中提出的算法不仅在速度上可以满足实际的应用,而且在保持检测边缘的RGB图像的视觉效果方面具有显著优越性.

The existing edge detection operator loses the original image visual effect during furniture image edge detection, so they are hardly used in furniture customization system. In order to solve this problem,this paper proposes a kind of Sobel edge detection method based on the new convolution kernel. Firstly, this method divides the RGB image into R image, G image, and B image in light of tricolor. Secondly,the convolution kernel described in the paper is used to make discrete weighted calculation for divided images. Experiments show that the algorithm proposed not only meets the practical application in terms of speed, but also has remarkable superiority in maintaining the visual effect of RGB image edge detection.