视频提取关键帧工具类KeyFramesExtractUtils.py，动态支持三种取帧方式，关键参数可配置，代码经过优化处理，效果和性能更好。

同步上一期视频关键帧提取方法：

视频提取关键帧的三种方式【已调通】_君临天下tjm的博客-CSDN博客_视频关键帧提取关键代码如下：# -*- coding: utf-8 -*-"""this key frame extract algorithm is based on interframe difference.The principle is very simpleFirst, we load the video and compute the interframe difference between each framesThen, we can choose one of t...https://blog.csdn.net/shanxiderenheni/article/details/124499862

项目中可以直接导入该工具类KeyFramesExtractUtils.py，三种取帧方式通过参数选择，method可以取"use_top_order"、"use_thresh"、"use_local_maxima"中的任何一种，第三种方式效果最佳，默认也是"use_local_maxima"。

优化点：

key_frame = 10  # 所隔帧数
cap.set(cv2.CAP_PROP_POS_FRAMES, j)  # 这里的cv2.CAP_PROP_POS_FRAMES参数是说：取第j帧之后的那一帧
j += key_frame
......
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

视频提取关键帧工具类KeyFramesExtractUtils.py完整代码如下：

# -*- coding: utf-8 -*-
"""
this key frame extract algorithm is based on interframe difference.The principle is very simple
First, we load the video and compute the interframe difference between each framesThen, we can choose one of these three methods to extract keyframes, which are
all based on the difference method:1. use the difference orderThe first few frames with the largest average interframe differenceare considered to be key frames.
2. use the difference thresholdThe frames which the average interframe difference are large than thethreshold are considered to be key frames.
3. use local maximumThe frames which the average interframe difference are local maximum areconsidered to be key frames.It should be noted that smoothing the average difference value beforecalculating the local maximum can effectively remove noise to avoidrepeated extraction of frames of similar scenes.After a few experiment, the third method has a better key frame extraction effect.
"""
import cv2
import operator
import numpy as np
import matplotlib.pyplot as plt
import sys
import time
from scipy.signal import argrelextrema# USE_LOCAL_MAXIMA=True，效果还行，时间有点长，54张图，包括一张差异趋势图
# 使用cv2.CAP_PROP_POS_FRAMES间隔取帧，主要消耗CPU，效率比内存高，效果更好# 视频提取关键帧工具类
class KeyFramesExtractUtils:# 初始化def __init__(self, video_path=None, save_path=None):self.video_path = video_pathself.save_path = save_path# 提取关键帧def extract_keyframe(self, method="use_local_maxima"):print(sys.executable)print("method===>", method)# fixed threshold valuethresh = 0.6# Number of top sorted framesnum_top_frames = 50# smoothing window sizelen_window = int(50)print("target video: " + self.video_path)print("frame save directory: " + self.save_path)# load video and compute diff between framescap = cv2.VideoCapture(str(self.video_path))curr_frame = Noneprev_frame = Noneframe_diffs = []frames = []video_frames = []key_frame = 10  # 所隔帧数t0_start = time.time()k = 0j = 0while True:if key_frame >= 1:t1_start = time.time()print("j======>", j)cap.set(cv2.CAP_PROP_POS_FRAMES, j)  # 这里的cv2.CAP_PROP_POS_FRAMES参数是说：取第j帧之后的那一帧j += key_framesuccess, frame = cap.read()if not success:print("第一次视频帧读取完毕!")breakgray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)curr_frame = grayif curr_frame is not None and prev_frame is not None:# logic herediff = cv2.absdiff(curr_frame, prev_frame)diff_sum = np.sum(diff)diff_sum_mean = diff_sum / (diff.shape[0] * diff.shape[1])frame_diffs.append(diff_sum_mean)temp_frame = Frame(k, diff_sum_mean)frames.append(temp_frame)video_frames.append(frame)k = k + 1prev_frame = curr_framet1_end = time.time()print("计算一次差分耗时：", (t1_end - t1_start), "s")cap.release()t0_end = time.time()print("计算共耗时：", (t0_end - t0_start), "s")t_start = time.time()# compute keyframekeyframe_id_set = set()if method == "use_top_order":print("---------------Using use_top_order---------------")# sort the list in descending orderframes.sort(key=operator.attrgetter("diff"), reverse=True)for keyframe in frames[:num_top_frames]:keyframe_id_set.add(keyframe.id)elif method == "use_thresh":print("---------------Using Threshold---------------")for i in range(1, len(frames)):if rel_change(np.float(frames[i - 1].diff), np.float(frames[i].diff)) >= thresh:keyframe_id_set.add(frames[i].id)else:print("---------------Using Local Maxima---------------")diff_array = np.array(frame_diffs)sm_diff_array = smooth(diff_array, len_window)frame_indexes = np.asarray(argrelextrema(sm_diff_array, np.greater))[0]for i in frame_indexes:keyframe_id_set.add(frames[i - 1].id)plt.figure(figsize=(40, 20))plt.locator_params(numticks=100)plt.stem(sm_diff_array)plt.savefig(self.save_path + 'plot.png')for idx in keyframe_id_set:name = "keyframe_" + str(idx) + ".jpg"cv2.imwrite(self.save_path + name, video_frames[idx])t_end = time.time()print("取帧共耗时：", (t_end - t_start), "s")class Frame:"""class to hold information about each frame"""def __init__(self, id, diff):self.id = idself.diff = diffdef __lt__(self, other):if self.id == other.id:return self.id < other.idreturn self.id < other.iddef __gt__(self, other):return other.__lt__(self)def __eq__(self, other):return self.id == other.id and self.id == other.iddef __ne__(self, other):return not self.__eq__(other)def rel_change(a, b):x = (b - a) / max(a, b)print(x)return xdef smooth(x, window_len=13, window='hanning'):"""smooth the data using a window with requested size."""print(len(x), window_len)s = np.r_[2 * x[0] - x[window_len:1:-1],x, 2 * x[-1] - x[-1:-window_len:-1]]# print(len(s))if window == 'flat':  # moving averagew = np.ones(window_len, 'd')else:w = getattr(np, window)(window_len)y = np.convolve(w / w.sum(), s, mode='same')return y[window_len - 1:-window_len + 1]if __name__ == "__main__":keyFrame = KeyFramesExtractUtils(video_path="e956ed44fffe4bfb97cf23474fb48ef3.avi", save_path="./extract_result/")keyFrame.extract_keyframe(method="use_local_maxima")# 计算共耗时： 307.6547501087189 s
# 取帧共耗时： 22.484756231307983 s

提取的关键帧如下：

检测结果总结：视频帧的提取率约为1.66‰，21m15s的视频，耗时约330s（即5分30秒）。

视频提取关键帧工具类KeyFramesExtractUtils.py，动态支持三种取帧方式，关键参数可配置，代码经过优化处理，效果和性能更好。相关推荐

视频提取关键帧的三种方式【已调通】
推荐优化后的视频关键帧提取方法,已经包装成工具类,代码做了优化,性能和效果更好. 视频提取关键帧工具类KeyFramesExtractUtils.py,动态支持三种取帧方式,关键参数可配置,代码经过优 ...
Spring视频转码工具类ffmpeg
使用的是ffmepg,需要先下载,然后解压到指定文件夹, 其中的变量ffmpegPath指定的是ffmepg解压后的exe路径这里看到processVideoFormat()方法中,推荐一律使用Mg ...
3个简单好用的视频提取字幕工具，效率极高，建议收藏
分享3个简单且效率非常高的视频提取字幕工具,提取后的字幕不仅可以保存为文档,还能直接导出为字幕格式的文件,不管是识别中文还是英文都能一键将视频字幕提取成功. 1.网易见外工作台一个免费的在线各种音视 ...
python视频提取关键帧_如何使用Python提取视频的关键帧？
在很多场景下,我们不想或者不能处理视频的每一帧图片,这时我们希望能够从视频中提取出一些重要的帧进行处理,这个过程我们称为视频关键帧提取. 关键帧提取算法多种多样,如何实现主要取决于你对于关键帧的定义. ...
Python正则表达式工具类文件的封装实例，提供了多个实例，并且在代码中包含中文注释
Python正则表达式工具类文件的封装实例,提供了多个实例,并且在代码中包含中文注释 import reclass RegexUtils:'''正则表达式工具类'''def __init__(self ...
Javascript定义类（class）的三种方法
将近20年前,Javascript诞生的时候,只是一种简单的网页脚本语言.如果你忘了填写用户名,它就跳出一个警告. 如今,它变得几乎无所不能,从前端到后端,有着各种匪夷所思的用途.程序员用它完成越来越 ...
视频监控系统中的流媒体服务器,视频监控系统中的流媒体服务器、直写和全切换三种取流架构方案...
原标题:视频监控系统中的流媒体服务器.直写和全切换三种取流架构方案一.流媒体服务器架构前摄像头视频信号通过转发流媒体服务器转发至上壁面显示和终端接入,视频存储磁阵列通过流媒体存储服务器写入.实时流 ...
python视频提取关键帧_一种视频关键帧提取算法的制作方法
本发明属于信息安全技术领域,涉及视频内容信息的提取,具体来说,是一种视频关键帧提取算法. 背景技术: 随着Internet的应用和普及,多媒体信息检索系统对社会各领域产生越来越大的影响.传统的信息检索 ...
抖音视频解析的工具类
提示:文章写完后,目录可以自动生成,如何生成可参考右边的帮助文档一.一个解析抖音分析链接的工具类 import java.io.IOException; import java.util.HashM ...

视频提取关键帧工具类KeyFramesExtractUtils.py，动态支持三种取帧方式，关键参数可配置，代码经过优化处理，效果和性能更好。

视频提取关键帧工具类KeyFramesExtractUtils.py，动态支持三种取帧方式，关键参数可配置，代码经过优化处理，效果和性能更好。相关推荐

最新文章

热门文章