Unity中基于SDL获取显微镜视频流RGB原始数据并解析的问题

我通过SDL库获取到了显微镜的视频流frameData数据，但不知道如何解析这些数据，生成Unity可使用的视频内容。相关代码如下：

public class CameraStreamManager : MonoBehaviour
{
    private Thread cameraThread;
    private bool isRunning = false; 
    private SLcamNet.NetCamera _camera; 
    private Texture2D videoTexture;

    public RawImage videoImage; 

    private List<NetInfo> _infos;
    private CameraInfo _cameraInfo;
    private SDLPanel sdlPanel; 
    private Task _task; 

    private int width = 1920;  
    private int height = 1080;
    private bool frameReady = false;

    public void InitializeCamera()
    {
        SLcamNet.NetCamera.InitNet();
        _infos = SLcamNet.NetCamera.SearchAvailableCameras();
        if (_infos.Count < 1)
        {
            Thread.Sleep(100);
            _infos = SLcamNet.NetCamera.SearchAvailableCameras();
        }
    }

    public void OpenCamera()
    {
        sdlPanel?.Dispose();
        sdlPanel = new SDLPanel
        {
            Anchor = AnchorStyles.Top | AnchorStyles.Bottom | AnchorStyles.Left | AnchorStyles.Right
        };

        Debug.LogError("_infos>>>>  " + _infos.Count);
        if (_infos.Count < 1)
        {
            Thread.Sleep(100);
            _infos = SLcamNet.NetCamera.SearchAvailableCameras();
            return;
        }

        CameraSettings settings = new CameraSettings
        {
            netInfo = _infos[0],
            resolution = new CameraResolution(1920, 1080),
            streamFormat = StreamFormat.kMJPEG,
            pixFormat = PixFormat.kRGBA8888,
            rcMode = "cbr",
            quality = "high",
            frameRate = 25,
            bitrate = 86000,
            keyFrameInterval = 90
        };
        settings.netInfo.Username = "admin";
        settings.netInfo.Password = "123456";

        _camera = SLcamNet.NetCamera.Create(settings.netInfo.Model);
        _camera.SetCameraSettings(settings);
        _camera.Login();
        _cameraInfo = _camera.GetNetCameraAllInfo();
        _camera.Open();

        isRunning = true;
        _task = Task.Run(() => StartCameraCapture());
    }

    private void StartCameraCapture()
    {
        Task.Run(() =>
        {
            Thread.Sleep(500); 
            IntPtr frameDataPtr = IntPtr.Zero;

            while (isRunning)
            {
                try
                {
                    if (!_camera.GetVideoFrameData(ref frameDataPtr))
                        continue;

                    SLcamNet.l_md_data_t frameData = SLcamNet.l_md_data_t.FromIntPtr(frameDataPtr);
                    Debug.LogError("frameData>>>>>" + frameData.p_rgb);
                }
                catch (Exception ex)
                {
                    Debug.LogError($"Error in StartCameraCapture: {ex.Message}");
                }
            }
        });
    }

    private void OnApplicationQuit()
    {
        isRunning = false;
        _task = null;

        if (sdlPanel != null)
        {
            sdlPanel.Dispose(true);
            sdlPanel = null;
        }

        _camera?.Close();
    }
}

解析与显示实现方案 #

你的代码已经拿到了帧数据指针frameData.p_rgb，结合Unity的纹理系统，只需完成数据复制→主线程更新纹理两个核心步骤，即可实现实时视频显示：

1. 新增线程安全数据缓存变量 #

在类中添加用于缓存帧数据的线程安全变量：

private byte[] frameBuffer;
private object bufferLock = new object();

2. 初始化Texture2D #

在OpenCamera方法中，相机启动后初始化与分辨率匹配的Texture2D，并绑定到RawImage：

// 在_camera.Open()之后添加
videoTexture = new Texture2D(width, height, TextureFormat.RGBA32, false);
videoImage.texture = videoTexture;

3. 子线程复制帧数据 #

修改StartCameraCapture中的帧处理逻辑，将非托管指针的RGBA数据复制到托管数组，标记帧就绪状态：

private void StartCameraCapture()
{
    Task.Run(() =>
    {
        Thread.Sleep(500); 
        IntPtr frameDataPtr = IntPtr.Zero;

        while (isRunning)
        {
            try
            {
                if (!_camera.GetVideoFrameData(ref frameDataPtr))
                    continue;

                SLcamNet.l_md_data_t frameData = SLcamNet.l_md_data_t.FromIntPtr(frameDataPtr);
                int dataLength = width * height * 4; // RGBA8888每个像素4字节

                lock(bufferLock)
                {
                    if(frameBuffer == null || frameBuffer.Length != dataLength)
                    {
                        frameBuffer = new byte[dataLength];
                    }
                    // 从非托管内存复制数据到托管数组
                    System.Runtime.InteropServices.Marshal.Copy(frameData.p_rgb, frameBuffer, 0, dataLength);
                    frameReady = true;
                }

                // 关键：记得调用SDK的帧释放方法（根据SDK文档调整）
                // _camera.ReleaseVideoFrameData(frameDataPtr);
            }
            catch (Exception ex)
            {
                Debug.LogError($"Error in StartCameraCapture: {ex.Message}");
            }
        }
    });
}

4. 主线程更新纹理 #

添加Update方法，在主线程中检测帧就绪状态，更新Texture2D：

private void Update()
{
    if(frameReady)
    {
        lock(bufferLock)
        {
            videoTexture.LoadRawTextureData(frameBuffer);
            videoTexture.Apply();
            frameReady = false;
        }
    }
}

关键注意事项 #

• 线程安全：Unity的Texture操作必须在主线程执行，子线程仅负责数据复制，绝对不能直接修改Texture
• 内存释放：务必调用SDK提供的帧数据释放方法（注释中的ReleaseVideoFrameData），避免内存泄漏
• 格式匹配：确保TextureFormat.RGBA32与相机设置的PixFormat.kRGBA8888完全匹配，否则会出现画面错乱
• 分辨率适配：如果相机分辨率动态变化，需要重新创建Texture2D和frameBuffer

内容的提问来源于stack exchange，提问作者jindeus ye