前言 Windows下实现摄像视频捕捉有多种实现方式;各种方式的优劣,本文不做对比。但是,opencv是一款老牌开发库,在图像处理领域声名显赫。采用opencv来处理摄像视频,在性能和稳定性上,是有保障的。并且,opencv包含很多图像处理函数,可以更方便的对视频处理。
执行程序是用wpf开发的,所以先将opencv封装成c语言接口,以供调用。opencv也不可能提供现成的控件供wpf使用,两种不同的开发语言“沟通”起来有些困难。其实稍作变通,就可以实现摄像头播放功能。
1 对opencv封装
opencv的类VideoCapture封装了对摄像头的操作,使用起来也非常简单。
bool open(int device); device为摄像头设备序号。
如果有多个摄像头,怎么知道哪个摄像头的序号那?可以通过如下函数,获取摄像头列表。摄像头在list中索引即为设备序号。
int GetCameraDevices(vector<wstring>& list) { ICreateDevEnum *pDevEnum = NULL; IEnumMoniker *pEnum = NULL; int deviceCounter = 0; CoInitialize(NULL); HRESULT hr = CoCreateInstance(CLSID_SystemDeviceEnum, NULL, CLSCTX_INPROC_SERVER, IID_ICreateDevEnum, reinterpret_cast<void**>(&pDevEnum)); if (SUCCEEDED(hr)) { // Create an enumerator for the video capture category. hr = pDevEnum->CreateClassEnumerator( CLSID_VideoInputDeviceCategory, &pEnum, 0); if (hr == S_OK) { //if (!silent)printf("SETUP: Looking For Capture Devices\n"); IMoniker *pMoniker = NULL; while (pEnum->Next(1, &pMoniker, NULL) == S_OK) { IPropertyBag *pPropBag; hr = pMoniker->BindToStorage(0, 0, IID_IPropertyBag, (void**)(&pPropBag)); if (FAILED(hr)) { pMoniker->Release(); continue; // Skip this one, maybe the next one will work. } // Find the description or friendly name. VARIANT varName; VariantInit(&varName); hr = pPropBag->Read(L"Description", &varName, 0); if (FAILED(hr)) hr = pPropBag->Read(L"FriendlyName", &varName, 0); if (SUCCEEDED(hr)) { hr = pPropBag->Read(L"FriendlyName", &varName, 0); int count = 0; wstring str2 = varName.bstrVal; list.push_back(str2); } pPropBag->Release(); pPropBag = NULL; pMoniker->Release(); pMoniker = NULL; deviceCounter++; } pDevEnum->Release(); pDevEnum = NULL; pEnum->Release(); pEnum = NULL; } } return deviceCounter; }
总之,使用opencv打开摄像头非常简单。
打开之后,就是获取摄像头图像。视频其实就是图像的集合;每秒钟获取25幅图像,将其在控件上显示,就是视频。
Mat cameraImg; _pCapture >> cameraImg;
Mat类封装了对图像的操作。c#不可能操作Mat,需要将Mat中纯图像部分数据传递出来,图像才能被c#利用。
int Camera_GetImgData(INT64 handle, char* imgBuffer) { CameraInfo *pCameraInfo = (CameraInfo*)handle; Mat cameraImg; *(pCameraInfo->_pCapture) >> cameraImg; if (!cameraImg.empty()) { int height = cameraImg.rows;int dataLen = height * cameraImg.step; memcpy(imgBuffer, cameraImg.data, dataLen); return 0; } else { return 1; } }
cameraImg.data中存有图像数据,data的大小可以根据图像的高度、每行图像的步幅计算出来。c#调用此函数后,imgBuffer存放图像数据。对数据imgBuffer处理后,就可以在控件上显示。
c语言对opencv封装函数列表如下:
extern "C" { OpenCVCamera_API int Camera_GetCameraName(char* listName); OpenCVCamera_API INT64 Camera_CreateHandle(); OpenCVCamera_API void Camera_CloseHandle(INT64 handle); OpenCVCamera_API BOOL Camera_IsOpen(INT64 handle); OpenCVCamera_API int Camera_Open(INT64 handle, int index); OpenCVCamera_API int Camera_Close(INT64 handle); OpenCVCamera_API int Camera_GetImgInfo(INT64 handle,int& width,int& height,int& channel, int& step, int& depth); OpenCVCamera_API int Camera_GetImgData(INT64 handle, char* imgBuffer); //flipCode >0: 沿y-轴翻转, 0: 沿x-轴翻转, <0: x、y轴同时翻转 OpenCVCamera_API int Camera_GetImgData_Flip(INT64 handle, char* imgBuffer, int flipCode); OpenCVCamera_API int Camera_ImgData_Compress(int rows, int cols, int type, void* imgBuffer, int param,void* destBuffer,int* destLen); }
函数定义如下
class CameraInfo { public: VideoCapture * _pCapture; CameraInfo() { _pCapture = NULL; } ~CameraInfo() { if (_pCapture != NULL) delete _pCapture; } }; int Camera_GetCameraName(char* listName) { vector<wstring> listName2; GetCameraDevices(listName2); wstring all; for (size_t i = 0; i < listName2.size(); i++) { all += listName2[i]; all += L";"; } int n = all.length(); memcpy(listName, all.data(), n * 2); return listName2.size(); } INT64 Camera_CreateHandle() { CameraInfo *pCameraInfo = new CameraInfo(); return (INT64)pCameraInfo; } void Camera_CloseHandle(INT64 handle) { CameraInfo *pCameraInfo = (CameraInfo*)handle; delete pCameraInfo; } BOOL Camera_IsOpen(INT64 handle) { CameraInfo *pCameraInfo = (CameraInfo*)handle; if (pCameraInfo->_pCapture == NULL) return FALSE; BOOL open = pCameraInfo->_pCapture->isOpened(); return open; } int Camera_Open(INT64 handle, int index) { Camera_Close(handle); CameraInfo *pCameraInfo = (CameraInfo*)handle; pCameraInfo->_pCapture = new VideoCapture(); bool open = pCameraInfo->_pCapture->open(index); if (open) return 0; return 1; } int Camera_Close(INT64 handle) { CameraInfo *pCameraInfo = (CameraInfo*)handle; if (pCameraInfo->_pCapture == NULL) return 0; delete pCameraInfo->_pCapture; pCameraInfo->_pCapture = NULL; return 0; } //depth == enum{CV_8U=0,CV_8S=1,CV_16U=2,CV_16S=3,CV_32S=4,CV_32F=5,CV_64F=6} int Camera_GetImgInfo(INT64 handle, int& width, int& height, int& channel, int& step, int& depth) { CameraInfo *pCameraInfo = (CameraInfo*)handle; Mat cameraImg; *(pCameraInfo->_pCapture) >> cameraImg; if (!cameraImg.empty()) { height = cameraImg.rows; width = cameraImg.cols; channel = cameraImg.channels(); step = cameraImg.step; depth = cameraImg.depth(); return 0; } else { return 1; } } int Camera_GetImgData(INT64 handle, char* imgBuffer) { CameraInfo *pCameraInfo = (CameraInfo*)handle; Mat cameraImg; *(pCameraInfo->_pCapture) >> cameraImg; if (!cameraImg.empty()) { int height = cameraImg.rows; int width = cameraImg.cols; int channel = cameraImg.channels(); int dataLen = height * cameraImg.step; memcpy(imgBuffer, cameraImg.data, dataLen); return 0; } else { return 1; } } int Camera_GetImgData_Flip(INT64 handle, char* imgBuffer, int flipCode) { CameraInfo *pCameraInfo = (CameraInfo*)handle; Mat cameraImg; *(pCameraInfo->_pCapture) >> cameraImg; if (!cameraImg.empty()) { Mat dest; if (flipCode != 100) { flip(cameraImg, dest, flipCode); } else { dest = cameraImg; } int height = dest.rows; int width = dest.cols; int channel = dest.channels(); int dataLen = height * cameraImg.step; memcpy(imgBuffer, dest.data, dataLen); return 0; } else { return 1; } } BOOL MatToCompressData(const Mat& mat, std::vector<unsigned char>& buff) { if (mat.empty()) { return FALSE; } //提前分配内存 否则会报错 int n = mat.rows*mat.cols*mat.channels(); buff.reserve(n * 2); std::vector<int> param(2); param.push_back(CV_IMWRITE_JPEG_QUALITY); param.push_back(95);// default(95) 0-100 return cv::imencode(".jpg", mat, buff, param); } int Camera_ImgData_Compress(int rows, int cols, int type, void* imgBuffer, int param, void* destBuffer, int* destLen) { Mat src(rows, cols, type, imgBuffer); std::vector<unsigned char> buff; MatToCompressData(src, buff); if (*destLen < buff.size()) return -1; memcpy(destBuffer, buff.data(), buff.size()); *destLen = buff.size(); return 0; }
2 WPF实现视频播放
WPF的Image控件实现图像的显示。实现视频播放的逻辑为:设定一个定时器(时间间隔为40毫秒),每隔一段时间从opencv获取图像,在控件中显示。
<Image x:Name="imageVideoPlayer" Stretch="Uniform" ></Image>
实现图像显示代码
BitmapSource bitmapSource = _openCVCamera.GetBitmapSource(); if (bitmapSource == null) return false; imageVideoPlayer.Source = bitmapSource;
实现图像显示的关键是构建BitmapSource,暨:如何从opencv中获取图像数据构建BitmapSource。
//获取图像数据 if (!GetImgData(out byte[] imgData)) return null; //构建WriteableBitmap WriteableBitmap img = new WriteableBitmap(_imgWidth, _imgHeight, 96, 96, PixelFormats.Bgr24, null); img.WritePixels(new Int32Rect(0, 0, _imgWidth, _imgHeight), imgData, img.BackBufferStride, 0); img.Freeze();
至此,就可以显示摄像头图像了。
原文地址:https://www.cnblogs.com/yuanchenhui/p/11332728.html