preamble
Previously through the articleA few lines of code to implement Python to capture, play and save camera videos!Gives an overview of how to read, play and save videos, followed by the articlePython Mito technology is also just a few lines of code!Gives you an overview of how to make adjustments to the brightness, contrast, chroma, or sharpness of an image for basic image processing operations.
In fact, those two articles are both padding, both to introduce you to how to make picture quality enhancements to videos. In this article, we will combine the content of the previous two articles to talk about how to enhance the video picture quality.
If you want to see the results directly, you can pull them up at the end of the article.
principle
I don't know if you all played with this when you were a kid?
That's how the earliest animations were formed, and I remember when I was a kid there were these little books for sale.
In fact, the principle of video is the same, a video is composed of many pictures, a picture is a frame. So if we want to enhance the picture quality of the video, we can split it into operating on each frame of the picture, an operation we described earlier.
Therefore, the method of picture quality enhancement for videos can be divided into three steps: splitting-> processing-> synthesizing.
Implementation steps
broken up inseparate items
We talked about how to capture the video stream from the camera and how to read the video and play it back in the first article. No matter which method we go through, we are operating on frames. So what is called splitting here is getting every frame of the video stream we captured or read.
success, img1 = ()
# If the frame is read correctly, success is True
if not success:
break
('img1', img1)It's that simple and we can get every frame of the video.
deal with
After getting a frame of the video, we have to convert that frame into an image in a format that we can work with. Earlier, when we described how to do picture quality enhancement on images, we used theImageEnhance Methods related to this function, this function is inside the PIL image processing library, so we must read each of our frames into a format that PIL can handle:
image = (np.uint8(img1)) # Converted to a format that PIL can handle
After reading the image, we can enhance the image quality, we still use the code we talked about in the previous article:
# Image processing def img_enhance(image, brightness=1, color=1,contrast=1,sharpness=1): # Brightness enhancement enh_bri = (image) if brightness: image = enh_bri.enhance(brightness) # Chroma enhancement enh_col = (image) if color: image = enh_col.enhance(color) # Contrast enhancement enh_con = (image) if contrast: image = enh_con.enhance(contrast) # Sharpness enhancement enh_sha = (image) if sharpness: image = enh_sha.enhance(sharpness) return image
synthesize
After image processing, we need each frame to be composited to get our final video:
cap = ('Your video directory/xxx.mp4')
success, _ = ()
# Resolution-width
width = int((cv2.CAP_PROP_FRAME_WIDTH))
# Resolution-height
height = int((cv2.CAP_PROP_FRAME_HEIGHT))
# of total frames
frame_counter = int((cv2.CAP_PROP_FRAME_COUNT))
video_writer = ('Output.mp4', cv2.VideoWriter_fourcc('M', 'P', '4', 'V'), 15, (width, height), True)
while success:
success, img1 = ()
try:
image = (np.uint8(img1)) # Converted to a format that PIL can handle
img_enhanced = img_enhance(image, 2, 2, 2, 3)
video_writer.write((img_enhanced))
if (1) & 0xFF == ord('q'):
break
except:
break
()
video_writer.release()
()
I'm reading video in mp4 format here, so when compositing and writing video files, we need to use the
cv2.VideoWriter_fourcc('M', 'P', '4', 'V') This format.
I'm not modifying the resolution of the image here, I'm just getting the resolution of the original video separately, and then passing in the original resolution as a parameter when writing to the video file.
If you need to change the resolution of the video, you can use the following way:
(src, dsize[, dst[, fx[, fy[, interpolation]]]])
utilizationresize The method will be fine:
resized = (img, (width, height), interpolation = cv2.INTER_AREA)
effect
Let's take a look at the pre-processing video:
The video after processing (I'm processing rather randomly here, the parameters are written randomly) looks like this:
summarize
To this point, our video quality enhancement features are basically realized, the code is not complex, add up to this point. However, if you want to deal with the effect of their own satisfaction, or need to put some effort to adjust the parameters, to optimize. Even for each frame may be passed into the parameters are not the same, which requires you to slowly study.
Above is Python to achieve video quality enhancement of the sample code in detail, more information about Python video quality enhancement please pay attention to my other related articles!