我是PyQt的新手，正在尝试创建视频播放器。当用户截图某个帧时，我应该能够在当前帧之前检索5个帧，并在当前帧之后检索另外5个帧。到目前为止，我只能检索当前帧，而且很难确定要更改哪些部分代码以存储前一帧和后一帧。这是完整的代码：
# PyQt5 Video player
#!/usr/bin/env python
from PyQt5.QtCore import QDir, Qt, QUrl, pyqtSignal, QPoint, QRect, QObject
from PyQt5.QtMultimedia import QMediaContent, QMediaPlayer, QVideoFrame, QAbstractVideoSurface, QAbstractVideoBuffer, QVideoSurfaceFormat
from PyQt5.QtMultimediaWidgets import QVideoWidget
from PyQt5.QtWidgets import (QApplication, QFileDialog, QHBoxLayout, QLabel,
QPushButton, QSizePolicy, QSlider, QStyle, QVBoxLayout, QWidget)
from PyQt5.QtWidgets import QMainWindow,QWidget, QPushButton, QAction
from PyQt5.QtGui import QIcon, QPainter, QImage
import sys
import os
import os.path as osp
class VideoFrameGrabber(QAbstractVideoSurface):
frameAvailable = pyqtSignal(QImage)
def __init__(self, widget: QWidget, parent: QObject):
super().__init__(parent)
self.widget = widget
def supportedPixelFormats(self, handleType):
return [QVideoFrame.Format_ARGB32, QVideoFrame.Format_ARGB32_Premultiplied,
QVideoFrame.Format_RGB32, QVideoFrame.Format_RGB24, QVideoFrame.Format_RGB565,
QVideoFrame.Format_RGB555, QVideoFrame.Format_ARGB8565_Premultiplied,
QVideoFrame.Format_BGRA32, QVideoFrame.Format_BGRA32_Premultiplied, QVideoFrame.Format_BGR32,
QVideoFrame.Format_BGR24, QVideoFrame.Format_BGR565, QVideoFrame.Format_BGR555,
QVideoFrame.Format_BGRA5658_Premultiplied, QVideoFrame.Format_AYUV444,
QVideoFrame.Format_AYUV444_Premultiplied, QVideoFrame.Format_YUV444,
QVideoFrame.Format_YUV420P, QVideoFrame.Format_YV12, QVideoFrame.Format_UYVY,
QVideoFrame.Format_YUYV, QVideoFrame.Format_NV12, QVideoFrame.Format_NV21,
QVideoFrame.Format_IMC1, QVideoFrame.Format_IMC2, QVideoFrame.Format_IMC3,
QVideoFrame.Format_IMC4, QVideoFrame.Format_Y8, QVideoFrame.Format_Y16,
QVideoFrame.Format_Jpeg, QVideoFrame.Format_CameraRaw, QVideoFrame.Format_AdobeDng]
def isFormatSupported(self, format):
imageFormat = QVideoFrame.imageFormatFromPixelFormat(format.pixelFormat())
size = format.frameSize()
return imageFormat != QImage.Format_Invalid and not size.isEmpty() and \
format.handleType() == QAbstractVideoBuffer.NoHandle
def start(self, format: QVideoSurfaceFormat):
imageFormat = QVideoFrame.imageFormatFromPixelFormat(format.pixelFormat())
size = format.frameSize()
if imageFormat != QImage.Format_Invalid and not size.isEmpty():
self.imageFormat = imageFormat
self.imageSize = size
self.sourceRect = format.viewport()
super().start(format)
self.widget.updateGeometry()
self.updateVideoRect()
return True
else:
return False
def stop(self):
self.currentFrame = QVideoFrame()
self.targetRect = QRect()
super().stop()
self.widget.update()
def present(self, frame):
if frame.isValid():
cloneFrame = QVideoFrame(frame)
cloneFrame.map(QAbstractVideoBuffer.ReadOnly)
image = QImage(cloneFrame.bits(), cloneFrame.width(), cloneFrame.height(),
QVideoFrame.imageFormatFromPixelFormat(cloneFrame.pixelFormat()))
self.frameAvailable.emit(image) # this is very important
cloneFrame.unmap()
if self.surfaceFormat().pixelFormat() != frame.pixelFormat() or \
self.surfaceFormat().frameSize() != frame.size():
self.setError(QAbstractVideoSurface.IncorrectFormatError)
self.stop()
return False
else:
self.currentFrame = frame
self.widget.repaint(self.targetRect)
return True
def updateVideoRect(self):
size = self.surfaceFormat().sizeHint()
size.scale(self.widget.size().boundedTo(size), Qt.KeepAspectRatio)
self.targetRect = QRect(QPoint(0, 0), size)
self.targetRect.moveCenter(self.widget.rect().center())
def paint(self, painter):
if self.currentFrame.map(QAbstractVideoBuffer.ReadOnly):
oldTransform = self.painter.transform()
if self.surfaceFormat().scanLineDirection() == QVideoSurfaceFormat.BottomToTop:
self.painter.scale(1, -1)
self.painter.translate(0, -self.widget.height())
image = QImage(self.currentFrame.bits(), self.currentFrame.width(), self.currentFrame.height(),
self.currentFrame.bytesPerLine(), self.imageFormat)
self.painter.drawImage(self.targetRect, image, self.sourceRect)
self.painter.setTransform(oldTransform)
self.currentFrame.unmap()
class VideoWindow(QMainWindow):
def __init__(self, parent=None):
super(VideoWindow, self).__init__(parent)
self.setWindowTitle("PyQt Video Player Widget")
self.counter = 0
self.mediaPlayer = QMediaPlayer(None, QMediaPlayer.VideoSurface)
self.videoWidget = QVideoWidget()
self.videoFrame = QVideoFrame()
self.playButton = QPushButton()
self.playButton.setEnabled(False)
self.playButton.setIcon(self.style().standardIcon(QStyle.SP_MediaPlay))
self.playButton.clicked.connect(self.play)
self.positionSlider = QSlider(Qt.Horizontal)
self.positionSlider.setRange(0, 0)
self.positionSlider.sliderMoved.connect(self.setPosition)
self.errorLabel = QLabel()
self.errorLabel.setSizePolicy(QSizePolicy.Preferred,
QSizePolicy.Maximum)
# Create new action
openAction = QAction(QIcon('open.png'), '&Open', self)
openAction.setShortcut('Ctrl+O')
openAction.setStatusTip('Open video')
openAction.triggered.connect(self.openFile)
# Create exit action
exitAction = QAction(QIcon('quit.png'), '&Exit', self)
exitAction.setShortcut('Ctrl+Q')
exitAction.setStatusTip('Exit application')
exitAction.triggered.connect(self.exitCall)
screenshotAction = QAction(QIcon('screenshot.png'), '&Screenshot', self)
screenshotAction.setShortcut('Ctrl+S')
screenshotAction.setStatusTip('Screenshot scenes')
screenshotAction.triggered.connect(self.screenshotCall)
# Create menu bar and add action
menuBar = self.menuBar()
fileMenu = menuBar.addMenu('&File')
#fileMenu.addAction(newAction)
fileMenu.addAction(openAction)
fileMenu.addAction(screenshotAction)
fileMenu.addAction(exitAction)
# Create a widget for window contents
wid = QWidget(self)
self.setCentralWidget(wid)
# Create layouts to place inside widget
controlLayout = QHBoxLayout()
controlLayout.setContentsMargins(0, 0, 0, 0)
controlLayout.addWidget(self.playButton)
controlLayout.addWidget(self.positionSlider)
layout = QVBoxLayout()
layout.addWidget(self.videoWidget)
layout.addLayout(controlLayout)
layout.addWidget(self.errorLabel)
# Set widget to contain window contents
wid.setLayout(layout)
self.mediaPlayer.setVideoOutput(self.videoWidget)
self.mediaPlayer.stateChanged.connect(self.mediaStateChanged)
self.mediaPlayer.positionChanged.connect(self.positionChanged)
self.mediaPlayer.durationChanged.connect(self.durationChanged)
self.mediaPlayer.error.connect(self.handleError)
def openFile(self):
fileName, _ = QFileDialog.getOpenFileName(self, "Open Movie",
QDir.homePath())
self.path = osp.dirname(str(fileName))
if fileName != '':
self.mediaPlayer.setMedia(
QMediaContent(QUrl.fromLocalFile(fileName)))
self.playButton.setEnabled(True)
def exitCall(self):
sys.exit(app.exec_())
def screenshotCall(self):
#Call video frame grabber
self.grabber = VideoFrameGrabber(self.videoWidget, self)
self.mediaPlayer.setVideoOutput(self.grabber)
self.mediaPlayer.pause()
self.grabber.frameAvailable.connect(self.process_frame)
self.errorLabel.setText("Taking a screenshot of image "+str(self.counter)+" ....")
self.mediaPlayer.play()
self.mediaPlayer.setVideoOutput(self.videoWidget)
def play(self):
if self.mediaPlayer.state() == QMediaPlayer.PlayingState:
self.mediaPlayer.pause()
else:
self.mediaPlayer.play()
def mediaStateChanged(self, state):
if self.mediaPlayer.state() == QMediaPlayer.PlayingState:
self.playButton.setIcon(
self.style().standardIcon(QStyle.SP_MediaPause))
else:
self.playButton.setIcon(
self.style().standardIcon(QStyle.SP_MediaPlay))
def positionChanged(self, position):
self.positionSlider.setValue(position)
def durationChanged(self, duration):
self.positionSlider.setRange(0, duration)
def setPosition(self, position):
self.mediaPlayer.setPosition(position)
def handleError(self):
self.playButton.setEnabled(False)
self.errorLabel.setText("Error: " + self.mediaPlayer.errorString())
def process_frame(self, image):
# Save image here
filename = "screenshot" + str(self.counter).zfill(6)
self.path = 'C:/Desktop/temp'
image.save(self.path+'/{}.png'.format(str(filename)))
self.counter = self.counter+1
if __name__ == '__main__':
app = QApplication(sys.argv)
player = VideoWindow()
player.resize(720, 480)
player.show()
sys.exit(app.exec_())
我正在考虑检索当前帧号，并创建循环以检索前后帧，然后将其发送到process_frame以进行保存，但是由于目前为止，我不确定如何实现，我仍然很难理解视频采集卡背后的原理。
答案 0 :(得分：2)
您有一些选择。
首先是使用QMediaPlayer.setPosition将视频搜索到所需的帧，然后使用已经实现的抓取器抓取当前帧。然后，您将恢复职位。
实际上，这很难实现，因为您需要知道准确的帧时间，而PyQT并未将其作为其API的一部分。
要获取帧时间，可以使用FFMS2库及其Python bindings：
import ffms
source_file = "test/x264.mkv"
vsource = ffms.VideoSource(source_file)
print("Frames:", vsource.properties.NumFrames)
print("Times:", vsource.track.timecodes)
这是另一个解决方案的地方。如果仍然使用FFMS2，为什么还要使用媒体播放器来获取帧呢？您可以直接使用FFMS2：
>>> frame = vsource.get_frame(0)
>>> frame.EncodedWidth, frame.EncodedHeight
(128, 72)
>>> frame.planes[0]
array([16, 16, 16, ..., 16, 16, 16], dtype=uint8)
您输入帧号并收到一个numpy数组，可以将其转换为带有库的实际图像，例如PIL或scikit-image。因此，您只需要将当前流的位置转换为帧号。为此，您可以查询vsource.track.timecodes变量并找到最接近的帧编号的索引。要快速完成此操作，可以使用分而治之的技术（内置的Python bisect模块）。
请注意，选择视频时，您应该加载一次ffms.VideoSource，因为索引时间和帧信息可能会花费一些时间，并且您不希望用户在用户界面滞后选择做一个截图。为了进一步加快处理速度，如果需要经常重新打开文件，还可以缓存索引信息。 FFMS2文档中介绍了手动索引。