基于B-样条插值的图像边缘检测

为了能高质量地进行图像边缘检测,提出了一种新的基于B-样条插值的边缘检测方法.该方法为了保证图像边缘检测后的质量,采用3次B-样条对图像的像素点灰度值进行插值运算,提出了基于B-样条插值的一阶和二阶微分算子,并引入数字滤波技术加以实现.利用B-样条可分离的特性,实现了2维空间的基于B-样条插值的连续图像的重构函数,在此基础上提出了用于边缘检测的基于3次B-样条的一阶梯度模板和二阶微分模板.该方法非常适合于采用硬件的并行化实现,极大地提高了图像处理速度和边缘检测质量.

Image Edge Detection Based on B-Spline Interpolation

In order to realize high-quality edge detection for the image,this paper introduces a new approach for edge detection based on B-spline interpolation.To ensure high-quality of the edge detection image,the cubic B-spline interpolation is used to calculate the gray of pixels of the image,then the one-order difference operator and the two-order difference operator are designed based on the cubic B-spline interpolation.The new approach is realized by the digital filter technology.Considering the separability of B-spline,the continuous synthesis function of 2D space is realized based on B-spline interpolation.Then one-order gradient template and two-order differential template are used to the edge detection.By using the characteristic of parallelism and pipeline operation of FPGA,the new approach for the edge detection image is achieved by hardware implementation.The approach is suitable for realtime image processing,thus greatly accelerating image processing and improving effect of image edge detection.