I don't know if you played this when you were a child ？
That's how the earliest animation was formed , I remember when I was a child, there were such small books for sale .
In fact, the principle of video is the same , A video is made up of many pictures , A picture is a frame . So we need to enhance the image quality of the video , Can be split into pairs
Each frame of the picture to operate , We introduced this operation earlier .
therefore , Video quality can be enhanced in three steps ： Split -> Handle -> synthesis .
In the first article, we talked about how to capture the video stream of the camera , And how to read and play the video . Either way , We all operate through frames . the
The so-called splitting here is to obtain each frame of the video stream we captured or read .
success, img1 = cap.read()
# If the frame is read correctly ,success by True
if not success:
break
cv2.imshow('img1', img1)
It's that simple , We can get every frame of the video .
After getting a frame of the video , We're going to convert this frame into a picture in a format we can handle . When we introduced how to enhance the image quality
Hou , It uses ImageEnhance Related methods of this function , This function is PIL In the image processing library , So we have to read the pictures of each frame
become PIL Format that can be processed ：
image = Image.fromarray(np.uint8(img1)) # convert to PIL Format that can be processed
After reading the image , We can enhance the image quality , Here we still use the code mentioned in our last article ：
# The image processing
python Exchange of learning Q Group ：906715085####
def img_enhance(image, brightness=1, color=1,contrast=1,sharpness=1):
# Brightness enhancement
enh_bri = ImageEnhance.Brightness(image)
if brightness:
image = enh_bri.enhance(brightness)
# Chroma enhancement
enh_col = ImageEnhance.Color(image)
if color:
image = enh_col.enhance(color)
# Contrast enhancement
enh_con = ImageEnhance.Contrast(image)
if contrast:
image = enh_con.enhance(contrast)
# Sharpness enhancement
enh_sha = ImageEnhance.Sharpness(image)
if sharpness:
image = enh_sha.enhance(sharpness)
return image
Image processing finished , We need to synthesize each frame of image , To get our final video ：
cap = cv2.VideoCapture(' Your video directory /xxx.mp4')
success, _ = cap.read()
# The resolution of the - Width
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
# The resolution of the - Height
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
# The total number of frames
frame_counter = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
video_writer = cv2.VideoWriter(' Output .mp4', cv2.VideoWriter_fourcc('M', 'P', '4', 'V'), 15, (width, height), True)
while success:
success, img1 = cap.read()
try:
image = Image.fromarray(np.uint8(img1)) # convert to PIL Format that can be processed
img_enhanced = img_enhance(image, 2, 2, 2, 3)
video_writer.write(np.asarray(img_enhanced))
if cv2.waitKey(1) & 0xFF == ord('q'):
break
except:
break
cap.release()
video_writer.release()
cv2.destroyAllWindows()
What I read here is mp4 Video format , So when composing and writing video files , We need to use
cv2.VideoWriter_fourcc('M', 'P', '4', 'V') This format .
I didn't change the resolution of the picture here , Just get the resolution of the original video separately , Then when you write a video file , Pass in the original resolution as a parameter .
If you need to change the resolution of the video , You can use the following methods ：
cv2.resize(src, dsize[, dst[, fx[, fy[, interpolation]]]])
Use resize The method is ok ：
resized = cv2.resize(img, (width, height), interpolation = cv2.INTER_AREA)
I won't send out the effect here , Interested friends can try . Only this and nothing more , Our video image quality enhancement function is basically realized , No code
complex , That's all it adds up to . however , If you want to deal with it to your satisfaction , We still need to work hard to adjust the parameters , To optimize . Even for each frame it may be
The parameters passed in are different , This requires you to study it slowly . If you think it's ok , Please don't be stingy of your praise .
