耿春明, 方菲. 融合区域生长与霍夫变换的内窥图像分割算法[J]. 应用光学, 2014, 35(6): 1009-1015.
引用本文: 耿春明, 方菲. 融合区域生长与霍夫变换的内窥图像分割算法[J]. 应用光学, 2014, 35(6): 1009-1015.
Geng Chun-ming, Fang Fei. Endoscopic image segmentation method combined region growing and Hough transformation[J]. Journal of Applied Optics, 2014, 35(6): 1009-1015.
Citation: Geng Chun-ming, Fang Fei. Endoscopic image segmentation method combined region growing and Hough transformation[J]. Journal of Applied Optics, 2014, 35(6): 1009-1015.

融合区域生长与霍夫变换的内窥图像分割算法

Endoscopic image segmentation method combined region growing and Hough transformation

  • 摘要: 为准确地划分出实际内窥图像的有效检测区域,依据此类图像的具体特点提出一种综合区域生长和霍夫变换的分割算法。利用区域生长大致分割出感兴趣区域,可能会存在漏检边缘或虚假边缘,通过二值形态学处理对图像进行平滑滤波和去噪,采用Canny算子在抑制噪声的同时进行边缘检测,应用霍夫变换检测圆的算法确定图像内有效区域的位置。通过对90组实际内窥图像在Visual C++ 6.0上进行仿真,实验结果表明:有88组内窥图像能够精确地分割强光干扰且划分出有效检测区域;仅有2组图像分割出的强光干扰及划分出的有效检测区域不够准确。

     

    Abstract: An image segmentation method for effective detection area based on region growing and Hough transformation was proposed in view of the features of actual endoscopic image. Firstly, region growth was used to segment the region of interest (ROI) roughly, there might be missed edge or false edges. Then the image was smoothed to remove noise by morphological processing. Next, the Canny operator was used to detect the edge and suppress the noise at the same time . In the end, Hough transformation, which was applied to detect circles, was conducted to determine the image position of the effective regions. The simulations of 90 groups of actual endoscopic images with Visual C++ 6.0 were done.The experimental result shows that the method can segment the strong light interference and extract the effective detection regions accurately of 88 groups;only 2 of them were not accurate enough.

     

/

返回文章
返回