android wifi直接直播摄像头视频流

Question

我已经在两个android设备之间建立了wifi直接p2p连接，以720x480的分辨率将实时摄像头从设备a流式传输到设备B。它工作正常，但即使在近距离（<1米）也会有很大的波动。有时它是~15fps，然后在几秒钟内它会下降到~3fps（只是一个猜测）。基本功能是PreviewCallback的OnPreviewFrame中的一个可运行线程，该线程使用YuvImage（）将预览帧压缩为JPEG并将其写入OutputStream。

我的问题是：有没有更有效的方法来做到这一点？我不需要惊人的帧速率（可能是20…？）。它只需要更稳定一点。

        private PreviewCallback previewCb_ = new PreviewCallback() {
    public void onPreviewFrame(byte[] data, Camera c) {
        frame = data;
        imageFormat = c.getParameters().getPreviewFormat();
        if (!socket.isClosed()) {
            mHandler.post(new Runnable() {
                public void run() {
                    if (stream != null){
                        try
                        {
                            //Log.d(ChooseFunction.TAG, "writing to stream");
                            buffer.reset();
                            synchronized(frame){
                                new YuvImage(frame, imageFormat, CameraView.IMG_WIDTH, CameraView.IMG_HEIGHT, null).compressToJpeg(area, 100, buffer);
                            }
                            buffer.flush();
                            // write the content header
                            stream.write(("--" + boundary + "rn" +
                                    "Content-type: image/jpgrn" +
                                    "Content-Length: " + buffer.size() +
                                    "rnrn").getBytes());

                            buffer.writeTo(stream);
                            stream.write("rnrn".getBytes());
                            stream.flush();
                        }
                        catch (IOException e)
                        {
                            Log.d(ChooseFunction.TAG, e.getMessage());
                        }
                    }
                }
            });
        }
    }
};

Answer 1

我使用OpenCV进行此转换，以提高将相机预览数据转换为位图的性能。

//Use openCV manager or openCV native libs directly in your project
        Log.i(TAG, "Trying to load OpenCV library");
    if (!OpenCVLoader.initAsync(OpenCVLoader.OPENCV_VERSION_3_2_0, this, mOpenCVCallBack)) {
        Log.e(TAG, "Cannot connect to OpenCV Manager");
    }
private BaseLoaderCallback mOpenCVCallBack = new BaseLoaderCallback(this) {
    @Override
    public void onManagerConnected(int status) {
        switch (status) {
            case LoaderCallbackInterface.SUCCESS: {
                //Init mat object
                mYuv = new Mat(480 + 480 / 2, 640, CvType.CV_8UC1);
                mRgb = new Mat();
                Log.i(TAG, "OpenCV loaded successfully");
                OpenCVLoaded = true;
            }
            break;
            default: {
                super.onManagerConnected(status);
            }
            break;
        }
    }
};
if (OpenCVLoaded) {
                //put camera preview data in mat
                mYuv.put(0, 0, instantPhotoData);
                Imgproc.cvtColor(mYuv, mRgb, Imgproc.COLOR_YUV420sp2RGB, 4);
                // convert to bitmap:
                final Bitmap rawBitmap = Bitmap.createBitmap(mRgb.cols(), mRgb.rows(), Bitmap.Config.ARGB_8888);
                Utils.matToBitmap(mRgb, rawBitmap);
                previewImageView.setImageBitmap(rawBitmap);
                    }
                });
            }