【Python】スクリーンキャプチャ（mp4, avi, gif出力）
説明
タイトルの通りです。選択した範囲をスクリーンキャプチャして、動画ファイル（mp4, avi, gif）として出力します。
プログラムを起動すると、まず現在の画面をキャプチャしたWindowが現れるので、マウスをドラッグして録画領域を選択します。
ドラッグが終わるとそのまま録画が開始されます。
録画を中断するには、コマンドプロンプト（？）上で何かキーを押してください。
ではまた。
プログラム
# coding: utf-8 # ************************* # https://stackoverflow.com/questions/49268120/why-cannot-imagegrab-grab-capture-the-whole-screen # to capture whole screen: # -> Goto Python installed directory and Right-click on python.exe # -> Properties -> Compatibility tab -> check 'Disable display scaling on high DPI settings'. # ************************* import cv2 import numpy as np from PIL import ImageGrab from PIL import Image import time import msvcrt isSetArea = False isMouseDown = False pos_1 = [0,0] pos_2 = [50,50] mouse_x = 0 mouse_y = 0 # ImageGrab.grab()で全画面をキャプチャするためのもの（高DPI設定を無効化） from ctypes import windll user32 = windll.user32 user32.SetProcessDPIAware() def callback(event, x, y, flags, param): global pos_1, pos_2, isSetArea, isMouseDown, mouse_x, mouse_y if event == cv2.EVENT_LBUTTONDOWN: isMouseDown = True pos_1[0] = int(x) pos_1[1] = int(y) if event == cv2.EVENT_LBUTTONUP: pos_2[0] = int(x) pos_2[1] = int(y) isSetArea = True if event == cv2.EVENT_MOUSEMOVE: mouse_x, mouse_y = x, y def get_record_area(): global pos_1, pos_2,isMouseDown, mouse_x, mouse_y cv2.namedWindow("test", cv2.WINDOW_NORMAL) cv2.setMouseCallback("test", callback) img_cv = np.asarray(ImageGrab.grab()) while isSetArea is False: img = cv2.cvtColor(img_cv, cv2.COLOR_BGR2RGB) img2 = cv2.resize(img, (int(np.shape(img)[1] * 0.5), int(np.shape(img)[0] * 0.5))) img3 = img2 if isMouseDown: img3 = cv2.rectangle(img3, tuple(pos_1), (mouse_x, mouse_y), (255, 255, 0), thickness=2) else: img3 = cv2.putText(img3, "please select record area", (0, 50), cv2.FONT_HERSHEY_PLAIN, 1.5, (255, 255, 255), 5, cv2.LINE_AA) cv2.imshow("test", img3) cv2.waitKey(1) pos_1[0], pos_1[1] = pos_1[0]*2, pos_1[1]*2 pos_2[0], pos_2[1] = pos_2[0]*2, pos_2[1]*2 cv2.destroyWindow("test") cv2.destroyAllWindows() def save_mp4(images, capSize, fps): # fourcc = cv2.VideoWriter_fourcc(*"H264") fourcc = cv2.VideoWriter_fourcc(*"MP4V") writer = cv2.VideoWriter("sample.mp4", fourcc, fps, capSize) for img in images: writer.write(img) writer.release() def save_avi(images, capSize, fps): print("FPS: {}".format(fps)) print("Capsize : ", capSize) print("video saving...") fourcc = cv2.VideoWriter_fourcc(*"XVID") writer = cv2.VideoWriter("sample.avi", fourcc, fps, capSize) for img in images: writer.write(img) writer.release() def save_gif(images, capSize, fps): print("FPS: {}".format(fps)) print("Capsize : ", capSize) print("video saving...") images_pillow = [] duration = int(1000. / fps) for image in images: images_pillow.append(Image.fromarray(cv2.cvtColor(image, cv2.COLOR_RGB2BGR))) images_pillow[0].save('sample.gif', save_all=True, append_images=images_pillow[1:], optimize=False, duration=duration, loop=0) def record(): count = 0 get_record_area() render_img_list = [] print("recording start....") print("録画を停止するには何かキーを押してください") print(f"position = ({pos_1}, {pos_2})") img_cv = np.asarray(ImageGrab.grab()) print(np.shape(img_cv)) start_time = time.time() while True: img_cv = np.asarray(ImageGrab.grab()) img = cv2.cvtColor(img_cv, cv2.COLOR_BGR2RGB) img_area = img[pos_1[0]:pos_2[1], pos_1[0]:pos_2[0]] # img_area = img[0:100,0:100] render_img_list.append(img_area) count += 1. if msvcrt.kbhit(): break # 何か文字が押されたら終了 width, height = pos_2[0]-pos_1[0], pos_2[1]-pos_1[1] fps = count / (time.time() - start_time) save_gif(render_img_list, (width, height), fps) # mp4出力の時：save_mp4(render_img_list, (width, height), fps) # avi出力の時：save_avi(render_img_list, (width, height), fps) record() input("何かキーを押してください.....")
