parser.add_argument("-v", help="visualise?", action="store_true") parser.add_argument("-s", help="save?", action="store_true") args = parser.parse_args() # Read image and kernel. # im = cv2.imread(args.i, cv2.IMREAD_GRAYSCALE) # Instantiate Demosaic_NN with the given bayer pattern code [args.p], # and run on the input bayer image. # dem = Demosaic_NN(args.p) im_BGR = dem.run(im) if args.m: im_mono = Convert.bgr2mono( im_BGR, R_weight=args.rw, G_weight=args.gw, B_weight=args.bw) # Display? # if args.v: cv2.imshow("input", im) cv2.imshow("demosaiced", im_BGR) if args.m: cv2.imshow("mono", im_mono) cv2.waitKey(0) # Save? # if args.s: cv2.imwrite("demosaiced.png", im_BGR)
outputs_titles.append("demosaiced") else: im_BGR = im_in # Perform AWB, if requested. # if options.a: awb = AWB_GrayWorld(options.aws, options.ail, options.aat) im_BGR = awb.run(im_BGR) outputs.append(np.copy(im_BGR)) outputs_titles.append("auto white balanced") # Convert to mono by taking a weighting of the channels. # im_mono = Convert.bgr2mono(im_BGR, R_weight=options.drw, G_weight=options.dgw, B_weight=options.dbw) outputs.append(np.copy(im_mono)) outputs_titles.append("monochromed") # Remove s&p noise from image, if requested. # if options.f: kernel = np.loadtxt(options.fk, delimiter=',') wm = Filter_WM(kernel) im_filtered = wm.run(im_mono) outputs.append(np.copy(im_filtered)) outputs_titles.append("filtered") # Display? #