import numpy as np import cv2, os def CVShow(img, title = 'unNamed', max_h = 950, max_w = 1800): img = np.array(img) cv2.namedWindow(title, cv2.WINDOW_KEEPRATIO) H,W = img.shape[:2] if H > max_h: ratio = max_h / H H, W = int(ratio * H), int(ratio * W) if W > max_w : ratio = max_w / W H, W = int(ratio * H), int(ratio * W) cv2.resizeWindow(title, W, H) cv2.imshow(title, img) # 自动适应图片大小的，不能缩放 key = cv2.waitKey(0) if key == ord('s'): # wait for key to write or exit cv2.imwrite(title+'.jpg', img) cv2.destroyAllWindows() return key def Access_e_ophtha_MA(rootpath): subdirs, imgpaths = os.listdir(rootpath), [] for subdir in subdirs: # 访问子文件夹 subdirpath = os.path.join(rootpath, subdir) imgnames = os.listdir(subdirpath) if 'Thumbs.db' in imgnames:imgnames.remove('Thumbs.db') for name in imgnames: # 访问子文件夹下的图片 imgpaths.append((os.path.join(subdirpath, name), name)) return imgpaths def enhance(img): clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8)) return clahe.apply(img) def Morph_Operate(img): r1 = cv2.morphologyEx(img, cv2.MORPH_OPEN, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5)), iterations=1) R1 = cv2.morphologyEx(r1, cv2.MORPH_CLOSE, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5)), iterations=1) r2 = cv2.morphologyEx(R1, cv2.MORPH_OPEN, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (11, 11)), iterations=1) R2 = cv2.morphologyEx(r2, cv2.MORPH_CLOSE, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (11, 11)), iterations=1) r3 = cv2.morphologyEx(R2, cv2.MORPH_OPEN, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (23, 23)), iterations=1) R3 = cv2.morphologyEx(r3, cv2.MORPH_CLOSE, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (23, 23)), iterations=1) return R3 def Vessels_Mask(f5): ret, f6 = cv2.threshold(f5, 15, 255, cv2.THRESH_BINARY) # 用来计算 mask # f7 = cv2.morphologyEx(f6, cv2.MORPH_OPEN, cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(3,3)), iterations=1) mask = np.ones(f5.shape[:2], dtype="uint8") * 255 _, contours, hierarchy = cv2.findContours(f6, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE) for cnt in contours: if cv2.contourArea(cnt) <= 200: cv2.drawContours(mask, [cnt], -1, 0, -1) im = cv2.bitwise_and(f5, f5, mask=mask) ret, fin = cv2.threshold(im, 15, 255, cv2.THRESH_BINARY_INV) newfin = cv2.erode(fin, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3)), iterations=1) return newfin def Vessels_Extract(newfin): fundus_eroded = cv2.bitwise_not(newfin) xmask = np.ones(fundus_eroded.shape[:2], dtype="uint8") * 255 _, xcontours, xhierarchy = cv2.findContours(fundus_eroded.copy(), cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE) for cnt in xcontours: shape = "unidentified" peri = cv2.arcLength(cnt, True) approx = cv2.approxPolyDP(cnt, 0.04 * peri, False) condition = len(approx) > 4 and cv2.contourArea(cnt) <= 3000 and cv2.contourArea(cnt) >= 100 shape = "circle" if condition else "veins" if shape == "circle": cv2.drawContours(xmask, [cnt], -1, 0, -1) finimage = cv2.bitwise_and(fundus_eroded, fundus_eroded, mask=xmask) blood_vessels = cv2.bitwise_not(finimage) return blood_vessels def Vessel_Segmentation_Demo(path): img = cv2.imread(path) _, green, _ = cv2.split(img) contrast_enhanced_green = enhance(green) morph_contrast_enhanced_green = Morph_Operate(contrast_enhanced_green) f4 = cv2.subtract(morph_contrast_enhanced_green, contrast_enhanced_green) f5 = enhance(f4) newfin = Vessels_Mask(f5) blood_vessels = Vessels_Extract(newfin) result1 = np.hstack([contrast_enhanced_green, morph_contrast_enhanced_green]) result2 = np.hstack([f4, f5]) result3 = np.hstack([newfin, blood_vessels]) return np.vstack([result1, result2, result3]) if __name__ == '__main__': e_ophtha_MA = r'J:\Image Projects\Microaneurysm Detection\Datasets\2013_e_ophtha\e_ophtha_MA' imgpaths = Access_e_ophtha_MA(e_ophtha_MA + '\MA - CutBlack') for path,name in imgpaths: result = Vessel_Segmentation_Demo(path) key = CVShow(result, name) if key == ord('b'): break
