python opencv 视频分帧，图片合成保存

代码没有问题，如果出现报错，请更换opencv版本，升级或降级
安装opencv pip install opencv-python
pip install natsort

视频分成图片

import cv2
vidcap = cv2.VideoCapture('test.mp4')
count = 0
while True:
    success,image = vidcap.read()
    cv2.imwrite("output/frame%d.jpg" % count, image)
    if cv2.waitKey(1) == 27:
        break
    count += 1

图片合成视频

（使用opencv合成不能添加音频，可以使用moviepy或ffmpeg，在处理完之后添加音频。也可以看我另一篇ffmpeg的博客，使用ffmpeg进行视频合成）

import cv2
import os
from natsort import ns, natsorted

img_file = 'dataset2/jpg/' # 图片路径
video_size = (1280,720) # 视频大小必须和图片大小相同，否则无法播放
#完成写入对象的创建，第一个参数是合成之后的视频的名称，第二个参数是可以使用的编码器，第三个参数是帧率即每秒钟展示多少张图片，第四个参数是图片大小信息
videowrite = cv2.VideoWriter('test.mp4',-1,5,video_size)
img_list=os.listdir(img_file)
img_list = natsorted(img_list,alg=ns.PATH) #对读取的路径，按照win排序方式，进行排序
for img_name in img_list:
    img = cv2.imread(img_file + img_name)
    #写入参数，参数是图片编码之前的数据
    videowrite.write(img)
print('end!')

保存视频，各种编码格式 + 定时录像！！

import cv2
import time
camera=cv2.VideoCapture('rtsp://admin:123456789.@192.168.2.43/MPEG-4/ch1/main/av_stream') # 网络摄像头，也可以使用本地0或1
video_size = (int(camera.get(cv2.CAP_PROP_FRAME_WIDTH)),int(camera.get(cv2.CAP_PROP_FRAME_HEIGHT))) # 获取摄像头w,h
fps = camera.get(cv2.CAP_PROP_FPS) # 获取摄像头帧率 ！！！！！！！！！！
# fps = 15 #                             或者自定义
# camera.set(3,1280) # 设置摄像头读取图像的宽度
# camera.set(4,720) # 高度，，，，不需要就不用设置
videowrite = cv2.VideoWriter('test.mp4',cv2.VideoWriter_fourcc(*'avc1'),fps,video_size) # 保存位置，编码，帧数，大小
# 0x00000021 h264编码
# cv2.VideoWriter_fourcc(*'mp4v') opencv官方版mp4,并不是h264编码
# cv2.VideoWriter_fourcc(*'XVID') opencv官方版avi，无压缩xvid编码
# cv2.VideoWriter_fourcc(*'avc1') openh264 mp4 h264编码，需要去官方
# https://github.com/cisco/openh264/releases下载openh264-1.8.0-win64.dll并添加环境变量

start = time.time() # 定义初始时间戳 ！！！！！！！！

while(True):
    ret,frame=camera.read()
    #写入参数，参数是图片编码之前的数据
    if time.time() - start > 60: # 录多长时间，以秒计算！！！！！！！
        videowrite = cv2.VideoWriter(str(time.time())+'.mp4',cv2.VideoWriter_fourcc(*'avc1'),fps,video_size) # 保存完继续保存
        start = time.time() # 重新计时
        # break # 保存完退出
    cv2.imshow('Dynamic',frame)
    videowrite.write(frame)
    #按下q键退出并保存
    if cv2.waitKey(1) & 0xff==ord('q'):
        break
camera.release()
cv2.destroyAllWindows()

图像裁剪、PIL/opencv/io/torch互转

# opencv
img.shape # 高度，宽度，色彩通道数
cutimg = img[y1:y2,x1:x2] # 裁剪
img_size = cv2.resize(img, (int(x / 2), int(y / 2))) # 指定尺寸缩放
img_size = cv2.resize(img, (0, 0), fx=0.25, fy=0.25, interpolation=cv.INTER_NEAREST) # 按比例缩放
# PIL
image.size # 宽度 ， 高度 
im.crop((x1, y1, x2, y2)) # 裁剪
image_size = image.resize((width, height),Image.ANTIALIAS) # 指定尺寸缩放

# io转opencv      opencv转io
cv_img = cv2.cvtColor(image_io_img , cv2.COLOR_RGB2BGR)
image_io_img = cv2.cvtColor(cv_img , cv2.COLOR_BGR2RGB)

# PIL转opencv     opencv转PIL
pil_img = Image.fromarray(cv2.cvtColor(cv_img, cv2.COLOR_BGR2RGB))
cv_img = cv2.cvtColor(np.asarray(pil_img), cv2.COLOR_RGB2BGR)

# 使用opencv代替torchvision.utils.save_image
grid = torchvision.utils.make_grid(tensor, normalize=True, range=(-1, 1))
np_img = grid.detach().cpu().mul(255.0).numpy().transpose((1,2,0))
cv_img = cv2.cvtColor(np_img, cv2.COLOR_RGB2BGR)

# torch转opencv
np_img = tensor.detach().cpu().mul(255.0).numpy().squeeze(0).transpose((1,2,0))
cv_img = cv2.cvtColor(np_img, cv2.COLOR_RGB2BGR)

# opencv 转 base64
cv_img = cv2.imencode('.jpg',image_np)[1]
base_img = "data:image/png;base64," + str(base64.b64encode(cv_img), "utf-8")
# base64 转 opencv
base_img = base64.b64decode(base64_code)
cv_img = cv2.imdecode(np.fromstring(base_img, np.uint8), cv2.COLOR_RGB2BGR)
# base64 转 PIL
base_img = base64.b64decode(base64_str)
base_img = BytesIO(base_img)
pil_img = Image.open(base_img)
# PIL 转 base64
img_buffer = BytesIO()
pil_img.save(img_buffer, format='JPEG')
base_img = base64.b64encode(img_buffer.getvalue()).decode()

读取路径下所有文件，并按win排序方法进行排序

    from natsort import ns, natsorted
    import os
    pathB = 'work/'
    imgB_list = os.listdir(pathB)
    imgB_list = natsorted(imgB_list,alg=ns.PATH) #对读取的路径进行排序
    for imgB in imgB_list:
        print(pathB+imgB)

PIL、opencv读取网络图片

from urllib.request import urlopen
import io
from PIL import Image
touxiang = 'https://y.jpg'
image_bytes = urlopen(touxiang).read()
data_stream = io.BytesIO(image_bytes)
touxiang = Image.open(data_stream).convert("RGB")

import requests
file = requests.get("https://www.baidu.com/1.png")
img = cv2.imdecode(np.fromstring(file.content, np.uint8), 1)

多线程读取摄像头并进行处理

import cv2
from threading import Thread
import time

class LoadStreams:  # multiple IP or RTSP cameras
    def __init__(self, sources='streams.txt'):
        if os.path.isfile(sources):
    		try:
        		with open(sources, 'r') as f:
            		sources = [x.strip() for x in f.read().strip().splitlines() if len(x.strip())]
    		except:
        		sources = [sources]
		else:
    		sources = [sources]
		n = len(sources)
        self.imgs = [None] * n
        for i, s in enumerate(sources):
            # Start the thread to read frames from the video stream
            print('%g/%g: %s... ' % (i + 1, n, s), end='')
            cap = cv2.VideoCapture(eval(s) if s.isnumeric() else s)
            assert cap.isOpened(), 'Failed to open %s' % s
            w = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
            h = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
            fps = cap.get(cv2.CAP_PROP_FPS) % 100
            _, self.imgs[i] = cap.read()  # guarantee first frame
            thread = Thread(target=self.update, args=([i, cap]), daemon=True)
            print(' success (%gx%g at %.2f FPS).' % (w, h, fps))
            thread.start()

    def update(self, index, cap):
        # Read next stream frame in a daemon thread
        n = 0
        while cap.isOpened():
            n += 1
            # _, self.imgs[index] = cap.read()
            cap.grab()
            if n == 4:  # read every 4th frame
                _, self.imgs[index] = cap.retrieve()
                n = 0
            time.sleep(0.01)  # wait time

    def __iter__(self):
        self.count = -1
        return self

    def __next__(self):
        self.count += 1
        img0 = self.imgs.copy()
        if cv2.waitKey(1) == ord('q'):  # q to quit
            cv2.destroyAllWindows()
            raise StopIteration

        return img0

    def __len__(self):
        return 0  # 1E12 frames = 32 streams at 30 FPS for 30 years


dataset = LoadStreams('rtsp://admin:123456789.@192.168.2.112') # 多线程读取摄像头，可改为读取txt文本

for frame_list in dataset:
    for frame in frame_list :
        print('耗时任务')
        cv2.imshow('Dynamic',frame)
        #按下q键退出并保存
        if cv2.waitKey(1) & 0xff==ord('q'):
            break

opencv PIL 贴图

pil_bg = Image.new(mode="RGBA", size=(width,height),color=(255,255,255)) #创建背景
small = Image.open(path).convert("RGBA") # 小图像
pil_bg.paste(small,(startLeft,startTop,startLeft+small.size[0],startTop+small.size[1]),mask=small.split()[3])