A method for extracting retina vessel image is provided. The method includes: a G-channel image of a color retina image is captured a Gabor filter is applied to transfer the G-channel image an image binarization is applied to filtered images for determining vessel position in the vessel image each pixel value in the same position of each of the filtered images is compared, and the pixel having maximum pixel value are assigned as composite pixels the composite pixels are combined to construct a composite retina image a noise removing step is performed to remove noise in the composite retina mage, then a required retina vessel image is obtained, in which the noise is corresponded to the region that are not vessel a dilation method is applied to the retina vessel image for enhancing the retina vessel image. Through the aforementioned method, the retina vessel image can be extracted precisely.本發明提供一種視網膜血管影像的攫取方法,藉由:取出彩色視網膜影像的G通道影像;以Gabor濾波器對G通道影像進行轉換;對濾波影像進行二值化步驟以判斷血管位置;由各濾波影像中,對位置相對應之各像素點進行逐一比對,並取出具最大值之像素點當成一合成像素點;以各合成像素點構成合成視網膜影像;執行去噪化步驟去除非屬血管影像的雜訊而得到所需視網膜血管影像;以及透過膨脹型態學方法對視網膜血管影像進行處理而得到增強視網膜血管影像等步驟,可準確取出視網膜血管影像以利於後續判讀。S101~S109‧‧‧步驟
