背景

好多開發者在對接大牛直播SDK（官方）的Windows平臺RTMP推送時，不熟悉攝像頭調用，實際上，攝像頭調用邏輯并不復雜，以下是大概流程：

首先調用我們sdk接口獲取攝像頭個數，調用接口是：GetVideoCaptureDeviceNumber()

接著需要獲取每個攝像頭的id和name，調用接口是：GetVideoCaptureDeviceInfo()

接下來針對某個指定的攝像頭，需要獲取它的分辨率列表個數,?調用接口是:GetVideoCaptureDeviceCapabilityNumber()

根據攝像頭id和分辨率的index,獲取實際的分辨率，調用接口:GetVideoCaptureDeviceCapability()

對的，比如一個攝像頭有四個分辨率?（1920*1080，1080*720，640*480，?320*240），?調用GetVideoCaptureDeviceCapabilityNumber?拿到的值就是4，想拿第一個index的分辨率，就傳入GetVideoCaptureDeviceCapability(id,?0,?capability)。

接口對接

以C#設計為例，接口設計如下：

/** 獲取攝像頭數量* * pNumer: 返回設備數量* * 成功返回 NT_ERC_OK*/[DllImport(@"SmartPublisherSDK.dll")]public static extern UInt32 NT_PB_GetVideoCaptureDeviceNumber(ref Int32 pNumer);/** 返回攝像頭設備信息* * device_index: 設備索引* device_name_utf8: 設備名稱，傳NULL將不返回名稱，如果不是NULL的話, device_name_length必須大于等于256, 返回utf8編碼的設備名稱* device_name_length: 設備名稱緩沖大小，如果device_name_utf8是NULL，則傳入0, 否則必須大于等于256* device_unique_id_utf8: 設備唯一ID, 傳NULL將不返回ID,如果不傳NULL的話，device_unique_id_length必須大于等于1024，返回utf8編碼的設備ID* device_unique_id_length: 設備唯一ID緩沖代銷, 如果是device_unique_id_utf8NULL, 則傳入0，否則必須大于等于1024* * 成功返回 NT_ERC_OK*/[DllImport(@"SmartPublisherSDK.dll")]public static extern UInt32 NT_PB_GetVideoCaptureDeviceInfo(Int32 device_index,[MarshalAs(UnmanagedType.LPStr)] StringBuilder device_name_utf8,UInt32 device_name_length,[MarshalAs(UnmanagedType.LPStr)] StringBuilder device_unique_id_utf8,UInt32 device_unique_id_length);/** 返回攝像頭能力數* * device_unique_id_utf8: 設備唯一id* capability_number: 返回的設備能力數* * 成功返回 NT_ERC_OK*/[DllImport(@"SmartPublisherSDK.dll")]public static extern UInt32 NT_PB_GetVideoCaptureDeviceCapabilityNumber([MarshalAs(UnmanagedType.LPStr)] String device_unique_id_utf8,ref Int32 capability_number);/** 返回攝像頭能力* * device_unique_id_utf8: 設備唯一id* device_capability_index: 設備能力索引* capability: 設備能力* * 成功返回 NT_ERC_OK*/[DllImport(@"SmartPublisherSDK.dll", EntryPoint = "NT_PB_GetVideoCaptureDeviceCapability", CallingConvention = CallingConvention.StdCall)]public static extern UInt32 NT_PB_GetVideoCaptureDeviceCapability([MarshalAs(UnmanagedType.LPStr)] String device_unique_id_utf8,Int32 device_capability_index,ref NT_PB_VideoCaptureCapability capability);

調用邏輯如下：

private void FillCameraInfo(){int device_number = 0;if (NTBaseCodeDefine.NT_ERC_OK != NTSmartPublisherSDK.NT_PB_GetVideoCaptureDeviceNumber(ref device_number)){return;}if (device_number < 1){return;}cameras_.Clear();for (int i = 0; i < device_number; ++i){CameraInfo info = new CameraInfo();info.capabilities_ = new List<NT_PB_VideoCaptureCapability>();StringBuilder name = new StringBuilder(256);StringBuilder id = new StringBuilder(1024);if (NTBaseCodeDefine.NT_ERC_OK != NTSmartPublisherSDK.NT_PB_GetVideoCaptureDeviceInfo(i,name, 256,id, 1024)){continue;}info.name_ = name.ToString();info.id_ = id.ToString();int capability_number = 0;if (NTBaseCodeDefine.NT_ERC_OK != NTSmartPublisherSDK.NT_PB_GetVideoCaptureDeviceCapabilityNumber(id.ToString(), ref capability_number)){continue;}bool is_failed = false;for (int j = 0; j < capability_number; ++j){NT_PB_VideoCaptureCapability capability = new NT_PB_VideoCaptureCapability();if (NTBaseCodeDefine.NT_ERC_OK != NTSmartPublisherSDK.NT_PB_GetVideoCaptureDeviceCapability(id.ToString(), j, ref capability)){is_failed = true;break;}info.capabilities_.Add(capability);}if (!is_failed){cameras_.Add(info);}}}

在此之前，需要設置是采集攝像頭，還是屏幕或者窗口：

/*定義Video源選項*/public enum NT_PB_E_VIDEO_OPTION : uint{NT_PB_E_VIDEO_OPTION_NO_VIDEO = 0x0,NT_PB_E_VIDEO_OPTION_SCREEN = 0x1, // 采集屏幕NT_PB_E_VIDEO_OPTION_CAMERA = 0x2, // 攝像頭采集NT_PB_E_VIDEO_OPTION_LAYER = 0x3, // 視頻合并，比如桌面疊加攝像頭等NT_PB_E_VIDEO_OPTION_ENCODED_DATA = 0x4, // 已經編碼的視頻數據，目前支持H264NT_PB_E_VIDEO_OPTION_WINDOW = 0x5, // 采集窗口}

更詳細的集成參考，參見：Windows平臺RTMP直播推送集成簡要說明_樂學吧-CSDN博客

總結

以上是生活随笔為你收集整理的Windows平台RTMP推送摄像头对接介绍的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。