基于加强梯度边缘检测算法和FFT的穿刺针定位方法

The Needle Localization Method Based on Gradient Edge Detection Algorithm and FFT

目的:针对CT引导介入图像中存在金属穿刺针伪影以及其他噪声,要准确提取穿刺针边缘和针尖的精确位置。方法:将加强梯度边缘检测算法和FFT结合,对图像进行滤波、增强、分割。首先采取高斯滤波器(GLPF)进行低通滤波,然后由罗盘算子、Canny算子提取边缘。通过取样针尖不同坐标计算针道斜率,据此计算针与水平方向夹角。为此训练了300幅CT引导穿刺图像。结果与结论 :在MATLAB平台上仿真验证,结果表明,罗盘边缘算子相对应进针角度时,强化了针的边缘,加强梯度边缘检测方法和FFT能有效抑制图像噪声,并能提高针尖角度计算精度。

Objective： In order to accurately and efficiently extract the needle tip and the precise location of the edge, the artifacts of metal needle and other noise of the CT images to guide intervention must be done. Methods： Combining the strengthened gradient edge detection algorithm with FFT, the images were filtered, enhanced and segmentalized. Firstly, with a Gaussian filter （GLPF） low-pass filter, using the compass operators and canny operator edge extraction. Coordinates calculated by sampling different needle tip slopes. There are 300 Computed Tomography Guided images that have been tested. Results and Conclusions： In the MATLAB simulation platform, the experiments show that the edge the compass operators is relative to the needle point, the reinforced edge of the needle, the strengthened gradient edge detection method and FFT can effectively suppress image noise, and it can improve the tip angle calculation precision.