一、利用numpy生成视频

import numpy as np
import cv2def array2video(videoName, w, h):"""使用numpy生成视频:param videoName: :param w: :param h: :return: """fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')videoWriter = cv2.VideoWriter(videoName, fourcc, 25, (w, h))frameCount = 0while True:img = np.ones((h, w, 3), np.uint8)*255cv2.putText(img, str(frameCount), (300, 300), cv2.FONT_HERSHEY_SIMPLEX, 2, (0, 255, 0), 3)cv2.imshow("frame", img)videoWriter.write(img)frameCount = frameCount + 1if cv2.waitKey(1) & 0xFF == 27:videoWriter.release()  # 关闭摄像头breakif __name__ == "__main__":array2video("mix.avi", 640, 480)

在转的过程中需要注意几点，否则会出现相应的视频文件不出现，出现了但是视频文件很小且打不开，可以打开但是视频的每一帧和预想的不一样：

  1. 首先需要注意的是视频解码格式，我在OpenCV官网上看到就是以('M', 'J', 'P', 'G')为例（其他解码格式也可以去尝试），当然也有简单的方法，不需要一个个的试，把编码的参数设置为-1（如下代码），有的系统弹出对话框显示该系统支持的编码格式（选择其中一个就行），有的系统会把支持的编码格式输出。不过这种方法不是对所有系统都有效的，看运气咯。同时后缀名一定要是.avi格式。可以用isOpened()方法查看是否真的创建了。解码格式的种类可以在这里查看。

    videoWriter = cv.VideoWriter('aa.avi', -1, 10, (1080, 640))
  2. 程序的最后一定要释放VideoWriter资源，否则最后的视频可能也会打不开，或者重启软件或者机器后才能打开该视频。 
    videoWriter.release()
  3. 注意创建的VideoWriter对象时，视频的宽和高一定要匹配好，否则也会导致视频文件很小且打不开（其实就是视频帧没有被写入）。注意下面的两句的宽和高位置对换： 
    img = np.ones((640, 1080, 3), np.uint8)*255
    videoWriter = cv.VideoWriter('bb.avi', fourcc, 10, (1080, 640))

最后，如果是自己用python的numpy创建的图片，一定要使用np.uint8类型（如下），否则最后的视频文件的播放画面就会像下面一样。（如果是先将其保存成一张一张的图片，然后在每一帧的读取转换成视频，就可以不做要求这样的要求）

img = np.ones((640, 1080, 3), np.uint8)*255

错误的画面

应该有的样子
 
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time        :2020/11/9 15:24
# @Author      :weiz
# @ProjectName :autoLabeling
# @File        :array2vide.py
# @Description :numpy生成视频
import numpy as np
import cv2
import randomdef array2video(videoName, w, h, steps, radius):fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')videoWriter = cv2.VideoWriter(videoName, fourcc, 25, (w, h))frameCount = 0startX = w + 400 // 2endX = - (400 // 2)centerX = w // 2centerY = h // 2x = endXy = 0r = 0while True:img = np.ones((h, w, 3), np.uint8)#cv2.putText(img, str(frameCount), (300, 300), cv2.FONT_HERSHEY_SIMPLEX, 2, (0, 255, 0), 3)if x <= endX:x = startXr = radius[random.randint(0, len(radius)-1)] // 2else:x = x - steps[random.randint(0, len(steps)-1)]y = centerY + random.randint(-1, 3)cv2.circle(img, (x, y), r, (255, 255, 255), -1)cv2.imshow("frame", img)videoWriter.write(img)frameCount = frameCount + 1if cv2.waitKey(1) & 0xFF == 27:videoWriter.release()  # 关闭摄像头breakradius = [205, 210, 215, 220, 225, 230, 235, 240, 245, 255, 260, 265, 270, 275, 280, 285, 290, 295, 310, 320, 330, 340, 350, 360, 370, 380, 390]
steps = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]
if __name__ == "__main__":array2video("mix.avi", 640, 480, steps, radius)

二、图片合成视频

import os
import cv2def images2video(imagesPath, videoName):""":param imagesPath::return:"""imagesPath = os.path.abspath(imagesPath)imagesList = os.listdir(imagesPath)imagesList.sort()if imagesList == []:print("The entered address has no content!")fps = 24size = cv2.imread(os.path.abspath(os.path.join(imagesPath, imagesList[0]))).shape[:2]size = (size[1], size[0])print(size)fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')video = cv2.VideoWriter(videoName, fourcc, fps, size)frame = 0for imageName in imagesList:frame = frame + 1path = os.path.abspath(os.path.join(imagesPath, imageName))img = cv2.imread(path)video.write(img)if frame % 100 == 0:print("The %d frame has been added" % frame)video.release()imagesPath = "E:\\pycharmProject\\YOLOv3\\results"  # 包含图片的文件夹
videoName = "16_detect.avi"
if __name__ == "__main__":images2video(imagesPath, videoName)

三、调用摄像头录制视频

# -*- coding: utf-8 -*-
# @Author     : weiz
# @Time       : 2020/4/8 上午10:08
# @FileName   : video.py
# @Software   : PyCharm
# @Description:
import cv2def recordVideo(camID, saveName):"""使用摄像头录制视频:param camID: 摄像头的ID:param saveName: 录制视频的名字:return:"""video = cv2.VideoCapture(camID)fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')fps = video.get(cv2.CAP_PROP_FPS)                   # 帧率size = (int(video.get(cv2.CAP_PROP_FRAME_WIDTH)), int(video.get(cv2.CAP_PROP_FRAME_HEIGHT)))out = cv2.VideoWriter(saveName, fourcc, fps, size)  # 视频存储while True:ret, img = video.read()if ret is False:out.release()cv2.destroyAllWindows()breakcv2.imshow("frame", img)out.write(img)if cv2.waitKey(1) & 0xFF == 27:out.release()cv2.destroyAllWindows()breakif __name__ == "__main__":recordVideo(0, "123.avi")
#include <opencv2/opencv.hpp>using namespace cv;void recordVideo()
{VideoCapture cam(0);if (!cam.isOpened()){std::cout << "cam open failed!" << std::endl;}VideoWriter video;int fps = cam.get(CAP_PROP_FPS);  //获取摄像机帧率if (fps <= 0)fps = 25;video.open("out.avi", VideoWriter::fourcc('M', 'J', 'P', 'G'),fps, Size(cam.get(CAP_PROP_FRAME_WIDTH), cam.get(CAP_PROP_FRAME_HEIGHT)));Mat img;for (;;){cam.read(img);if (img.empty())break;imshow("cam", img);//写入视频文件video.write(img);if (waitKey(5) == 'q') break;}video.release();
}

四、视频转图片

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time        :2020/10/12 13:33
# @Author      :weiz
# @ProjectName :yoloProcessedData
# @File        :video2images.py
# @Description :将视频切分成图片
import os
import cv2
import shutildef checkFileExist(fileName, isCreate=None):"""检测当前工作目录下fileName文件是否存在，并根据isCreate是否创建:param fileName:文件夹名字:param isCreate:是否创建:return:不存在返回false，存在或者创建返回true"""if isCreate == True:isCreate = Trueelse:isCreate = Falseif os.path.exists(fileName):return Trueelse:if isCreate:os.mkdir(fileName)print("the path:{} is builded!".format(fileName))return Trueelse:return Falsedef video2image(videoPath, imagePath, gap, isClearImagePath):"""将视频分割图片:param videoPath: 视频所在路径:param imagePath: 保存图片的路径:param gap: 每隔多少取一帧:param isClearImagePath: imagePath文件夹存在是否删除并重新创建:return: """if not checkFileExist(videoPath):print("The {} is not exist!".format(videoPath))returnif isClearImagePath:if checkFileExist(imagePath):shutil.rmtree(imagePath)  # 删除原来的图片库，以防混乱print("Old {} deleted!".format(imagePath))checkFileExist(imagePath, True)else:if not checkFileExist(imagePath, True):print("The {} is not exist!".format(imagePath))returnvideoNames = os.listdir(videoPath)for videoName in videoNames:name, suffix = videoName.split(".")imagePath_sub = imagePath + "/" + name  # imagePath的子目录os.mkdir(imagePath_sub)videoRoot = videoPath + "/" + videoName  # 视频的绝对路径cap = cv2.VideoCapture(videoRoot)frameNum = 1imageNumName = 0print(videoRoot)while (True):ret, frame = cap.read()if not ret:breakif frameNum % gap == 0:# 图片的保存路径savePath = imagePath_sub + "/" + name + "_" + "{:0>6d}".format(imageNumName) + ".png"imageNumName = imageNumName + 1cv2.imwrite(savePath, frame)if frameNum % (100*gap) == 0:print(frameNum)frameNum = frameNum + 1videoPath = "./videos"
imagePath = "./images"
gap = 10
isClearImagePath = True
if __name__ == "__main__":video2image(videoPath, imagePath, gap, isClearImagePath)

五、视频的横屏和竖屏转换

# eg: VSAndHS("./1.mp4", 0)
def VSAndHS(videoPath, rotation, savePath=None):"""竖屏和横屏互转(Turn between vertical and horizontal screens):param videoPath: 视频路径:param rotation: 旋转角度:1是顺时针旋转90度，0是逆时针旋转90度:param savePath: 保存路径:return:"""if rotation not in [0, 1]:rotation = 1if savePath == None:savePath = os.path.dirname(videoPath)savePath = os.path.join(savePath, os.path.basename(videoPath).split('.')[-2]+"_r"+".avi")cap = cv2.VideoCapture(videoPath)ret, frame = cap.read()trans_img = cv2.transpose(frame)rotation_img = cv2.flip(trans_img, rotation)fourcc = cv2.VideoWriter_fourcc(*'XVID')            # 视频存储的格式fps = cap.get(cv2.CAP_PROP_FPS)                     # 帧率# 视频的宽高size = frame.shape[:2]                              # img.shape: 高，宽，通道out = cv2.VideoWriter(savePath, fourcc, fps, size)  # 视频存储out.write(rotation_img)while True:ret, frame = cap.read()if not ret:breaktrans_img = cv2.transpose(frame)rotation_img = cv2.flip(trans_img, rotation)    # 获取转化后的图片out.write(rotation_img)cv2.imshow("src", frame)cv2.imshow("dest", rotation_img)cv2.waitKey(1)cap.release()out.release()

六、多个视频合成一个视频

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time        :2021/4/9 15:33
# @Author      :weiz
# @ProjectName :pycharmProject
# @File        :videos2video.py
# @Description :多视频合成一个视频
import os
import cv2
import numpy as np
from PIL import Image, ImageDraw, ImageFontdef plotImage(img, chinese, bbox):"""在图片中显示中文:param img::param chinese::param loc::return:"""# cv2.rectangle(img, (bbox[0], bbox[1]), (bbox[2], bbox[3]), (255, 0, 0), 2)pilimg = Image.fromarray(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))draw = ImageDraw.Draw(pilimg)font = ImageFont.truetype("simhei.ttf", 15, encoding="utf-8")draw.text([bbox[0], bbox[1] - 15], chinese, (255, 0, 0), font=font)chineseImg = cv2.cvtColor(np.array(pilimg), cv2.COLOR_RGB2BGR)return chineseImgdef videos2video(videosPath, name, repeat=1):"""多个视频转合成一个视频videosPath：视频路径repeat=1：每个视频重复次数，默认一次"""fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')videoWriter = cv2.VideoWriter(name, fourcc, 22, (640, 360))videoNameList = os.listdir(videosPath)for videoName in videoNameList:videoNamePath = os.path.join(videosPath, videoName)for rep in list(range(repeat)):video = cv2.VideoCapture(videoNamePath)while True:ret, img = video.read()if ret is False:cv2.destroyAllWindows()breakimg = plotImage(img, videoName.split('.')[0], [250, 25, 300, 50])  # 用于添加中文cv2.imshow("frame", img)videoWriter.write(img)cv2.waitKey(1)videoWriter.release()videosPath = "C:\\Users\\weiz\\Desktop\\videos"
if __name__ == "__main__":videos2video(videosPath, "abc.avi", 4)

七、根据按键保存图片或视频

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time        :2022/1/4 13:37
# @Author      :weiz
# @ProjectName :Sfm-python-master
# @File        :get_image_vide.py
# @Description :
import cv2
import datetime
import osdef get_image(camID):"""根据按键保存图片:param camID::return:"""if not os.path.exists("./save_images"):os.mkdir("./save_images")video = cv2.VideoCapture(camID)while True:ret, img = video.read()if ret is False:print("The camera cannot be turned on!")cv2.destroyAllWindows()breakcv2.imshow("frame", img)k = cv2.waitKey(1)if k == ord('s'):now_time = datetime.datetime.now()name = os.path.join("./save_images", str(now_time).replace(':', '-') + ".png")cv2.imwrite(name, img)print("The {} is save!".format(name))elif k == 27:cv2.destroyAllWindows()breakdef recordVideo(camID, saveName=None):"""根据开关按钮的开与关，录制视频:param camID: 摄像头的ID:param saveName: 录制视频的名字:return:"""if saveName == None:now_time = datetime.datetime.now()saveName = str(now_time).replace(':', '-') + ".avi"video = cv2.VideoCapture(camID)fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')fps = video.get(cv2.CAP_PROP_FPS)  # 帧率size = (int(video.get(cv2.CAP_PROP_FRAME_WIDTH)), int(video.get(cv2.CAP_PROP_FRAME_HEIGHT)))out = cv2.VideoWriter(saveName, fourcc, fps, size)  # 视频存储switch_save = 0while True:ret, img = video.read()if ret is False:out.release()cv2.destroyAllWindows()breakcv2.imshow("frame", img)if switch_save:out.write(img)k = cv2.waitKey(1)if k == ord('s'):switch_save = 1elif k == ord('e'):switch_save = 0elif k == 27:out.release()cv2.destroyAllWindows()breakcamID = 0
def main():# get_image(camID)recordVideo(camID)if __name__ == "__main__":main()

opencv视频和图片之间的处理相关推荐

  1. OpenCV 视频与图片序列相互转换 VideoWriter生成视频流

    OpenCV学习笔记(四十七)--VideoWriter生成视频流highgui http://blog.csdn.net/yang_xian521/article/details/7440190 标 ...

  2. html让视频跟图片之间无缝,图片、样式、音视频之间的无缝拼接，教你一个万能方法！...

    原标题:图片.样式.音视频之间的无缝拼接,教你一个万能方法! 大家好,我是老沙. 发现没有,几乎每个品牌的公众号都开始长图排版了,长图排版最常遇到的一个问题便是缝隙.去缝的教程老沙写过不少: 关于静态 ...

  3. python代码实现视频与图片之间的互相转换

    一.视频与图片 1.视频   视频其实就是若干张图片叠加起来,包含了图像的运行信息.   由于人眼识别的帧率是有限的,因此在单位时间内给人眼观看的图像超过一定数目的时候,给人的感觉就好像画面是运动的, ...

  4. C++ OpenCV视频摄像头图片读写操作

    总结一下C++语言下OpenCV的基本读写操作. 头文件导入和CMakeLists.txt #include<opencv2/opencv.hpp> #include<iostrea ...

  5. html让视频跟图片之间无缝,如何做到两个视频之间自然无缝衔接

    Vlog是一种很好的记录日常生活,表达自己想法的输出方式.但通常我们拍好素材之后会发现场景之间的转换太过生硬.那么,在前期拍摄的时候我们就要注意尽量使所有的场景符合自己想要表达的主题,片段之间跳跃性不 ...

  6. 视频到图片（每隔几帧保存一张图片）opencv实现

    如果想要自动截取一个视频的图片,又不想手动暂停截取,那可以试试用opencv自动截取(比如想要自动截取喜欢电影的经典台词与图片). 背景说明:一个小意外,录屏视频没有声音,想着把视频里的PPT截取出来 ...

  7. OpenCV：01图片视频的加载显示

    文章目录 **学习目标:** 图像与视频的加载显示 导入OpenCV的包`cv2` 窗口操作 计算按键的`ASCII`值 读取图片 用其他插件来读取(以`matplotlib`为例) 用`OpenCV ...

  8. python图片转视频加特效_使用Python opencv实现视频与图片的相互转换

    因为最近要经常转换数据集进行实验,因此记录一下. 1.视频转图片 即为将视频解析为一帧一帧的图片: import cv2 vc=cv2.VideoCapture("/home/hqd/Pyc ...

  9. 通过opencv实现将图片转换为视频

    背景: 为了提升用户的覆盖度,需要将书法图片转化成书法视频,且需要有浏览的效果,这里采用opencv先将图片转化成视频,在利用ffmpeg对视频进行处理,增加音乐.视频编码等. 思路: 1.openc ...

最新文章

  1. Visual Studio代码中的多个游标[关闭]
  2. hibernate同时插入两条相同数据（偶尔）原因调查
  3. 一次解决libgcc_s.so.1 must be installed for pthread_cancel to work的经历
  4. 实战04_redis-cluster集群搭建
  5. 模板消息 php实例,PHP微信模板消息操作示例
  6. windows 弹shell_Windows系统常用免费软件“红黑榜”
  7. select 组件-bug解决方案
  8. 学JAVA第十三天，方法、方法重载及构造函数
  9. 李开复给中国学生的四封信（转载）
  10. 栅格模型数据编码方式
  11. 影子系统、还原精灵、冰点还原优缺点比较
  12. win7系统rpc服务器不可用无法开机,win7系统RPC服务器不可用的解决方法
  13. MATLAB图像分割的GUI设计
  14. uniapp 打印小票
  15. TPYBoard中编译MicroPython并利用DfuSe经行固件烧录
  16. HDU2058 The sum problem（数学问题）
  17. Hantek6022BE 虚拟示波器 （二）方波 采样率 带宽
  18. 解决D-LINK DI 504路由器上网掉线问题
  19. Java高级：字节跳动历年校招Java面试真题解析
  20. 动态(视频)图像拼接和EDF景深融合解决方案（源码分享）

热门文章

  1. 如何免费获得短视频配乐
  2. 思科华为路由器如何利用route-map配置双wan口策略路由
  3. python安装包区别
  4. 微信账单页面明细格式化价格字符
  5. Java通过第三方工具实现Word、PPT转 PDF （easy）
  6. python中class什么意思_python中的class代表什么
  7. python 全部缩进一行_和你学Python之Python基础
  8. mysql单精度与双精度_单精度与双精度的区别
  9. SAP-MM发票校验过账到总账
  10. 奇安信2022秋招面经-技术支持岗