摘要
为了提取拉丝模具内孔的衍射图像的边缘信息,以往的检测方法都有各自的局限性,因此采用contourle tHMT变换结合LS—SVM的方法;首先利用contourlet HMT变换将图像分成粗糙系数和精细系数,通过LS—SVM训练粗糙系数找出边缘系数和噪声系数;利用BayesShrink方法去除噪声系数,采用梯度边缘系数和去噪后的系数,用加权平均值的方法进行图像的融合,最后使用contour—let HMT的逆变换对图像进行重构;对不同情况下的图像进行了测试,结果证明Contourlet HMT的边缘检测方法明显优于其它的方法。
In order to extract the edge information of the diffraction image of the inner hole in the wire drawing die.with its own limitations of conventional detection method,so the contourlet HMT transform is used combining LS-SVM.Firstly contourlet HMT divides an image into coarser coefficient and finer coefficient.Train the coarser coefficients and classify them using the LS-SVM classifier into edge and noise coefficient classes.Remove noise coefficient using the BayesShrink method and obtain a denoised coefficient class.Detect edge coefficient class and denoised coefficient class using gradient.Fuse edges information using weighted average fusion rule.Finally.reconstruct edge image using inverse contourlet HMT.The image is tested in different conditions.The results show that the edge detection method of HMT contourlet is better than other methods
引文
[1]孙珂琪,郑堤,陈俊华,等.图像检测螺纹的系统设计[J].计算机测量与控制,2015,23(3):893-896.
[2]李瑶,董瑞,何韬,等.一种基于contourlet变换的图像去噪方法[J].计算机技术与发展,2007,17(3):81-87.
[3]李杏梅.contourlet变换在图像去噪与边缘检测中的应用研究[D]华中科技大学,2011
[4]顾燕萍,赵文杰,吴占松.最小二乘支持向量机的算法研究[J].清华大学学报(自然科学版),2010,50(7):1063-1066.
[5]刘涵,郭勇,郑岗,等.基于最小二乘支持向量机的图像边缘检测研究[J].电子学报,2006,34(7):1275-1279.
[6]Gh.S.El-tawel,A.K.Helmy.An edge detection scheme based on least squares support vectormachine in a contourlet HMT domain[J].Applied Soft Computing 26(2015)418-427
[7]李万臣,赵开伟,陈宇贤.基于Contourlet变换在图像去噪中的应用[J].仪器仪表用户,2007,17(1):47-48.
[8]宋召青,郑苏,李志成.基于边缘检测与最小二乘支持向量机的人脸图像识别[J].海军航空工程学院学报,2010,25(5):518-522.