<分区>
out = cv2.VideoWriter(output_file, codec, fps, (width,height))
while video.isOpened():
has_frame, image = video.read()
if has_frame:
image_in = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image_in = tf.expand_dims(image_in, 0)
image_in = transform_images(image_in, FLAGS.size)
boxes, scores, classes, nums = get_yolo_model()(image_in)
image = draw_outputs(image, (boxes, scores, classes, nums), get_class_names())
image = cv2.putText(
image,
"",
(0, 30),
cv2.FONT_HERSHEY_COMPLEX_SMALL,
1,
(0, 0, 255),
2
)
out.write(image)
frame_count += 1
print('[In progress...] Writing frame:', frame_count)
else:
break
有没有什么方法可以引入线程的概念来更快地处理视频？
我试过使用这个:How to increase performance of OpenCV cv2.VideoCapture(0).read()
我尝试实例化 __init__ 方法线程，但是，如何将它与读取 while 循环内帧的 while 循环连接起来。