使用相机API使用渲染脚本中的RGB字节更新相机预览吗?

时间:2018-08-22 13:23:42

标签: java android image-processing android-ndk android-renderscript

我已经能够从我的相机预览中操纵字节数组,并使用renderscript从yuvtorgb更改字节。现在,我需要知道如何将rgb字节渲染回相机预览。我相信无法更新相机预览。我将需要在我的textureview顶部的某种叠加视图,然后将更新的字节数组呈现到叠加视图的表面。我需要一些指导我应该如何去做。有人可以建议我应该怎么做。这是我到目前为止尝试过的。

Mainactivity.java

public class MainActivity extends Activity implements TextureView.SurfaceTextureListener, Camera.PreviewCallback {
    private byte[] FrameData = null;

    private Camera mCamera;
    private TextureView mTextureView;
    Context mContext;
    private Camera.Size previewSize;
    private ImageView mimageView;

    private SurfaceTexture mTexture;
    private int[] pixels;

    private int pwidth;
    private int pheight;
    Renderscript rv;


    @Override
    protected void onCreate(Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        mimageView = new ImageView(this);
        mTextureView = new TextureView(this);
        mTextureView.setSurfaceTextureListener(this);

        rv = new Renderscript();

        setContentView(mTextureView);
    }

    @Override
    public void onSurfaceTextureAvailable(SurfaceTexture surface, int width, int height) {


        DisplayMetrics displayMetrics = this.getResources().getDisplayMetrics();
        int screenWidth = displayMetrics.widthPixels;
        int screenHeight = displayMetrics.heightPixels;

        int  mPreviewRate = Math.round((float)screenHeight / (float)screenWidth);
        mCamera = Camera.open();

        Camera.Size previewSize = mCamera.getParameters().getPreviewSize();

        mTextureView.setLayoutParams(new FrameLayout.LayoutParams(
                previewSize.width*mPreviewRate, previewSize.height*mPreviewRate, Gravity.CENTER));



        Camera.Parameters params = mCamera.getParameters();
        if (params.getSupportedFocusModes().contains(
                Camera.Parameters.FOCUS_MODE_CONTINUOUS_VIDEO)) {
            params.setFocusMode(Camera.Parameters.FOCUS_MODE_CONTINUOUS_VIDEO);
        }
        if (params.isAutoExposureLockSupported()) {
            params.setAutoExposureLock(false);
        }
        mCamera.setParameters(params);




        pwidth = previewSize.width*mPreviewRate;
        pheight = previewSize.height*mPreviewRate;

        try {
            mCamera.setPreviewTexture(surface);
        } catch (IOException t) {
        }
        mCamera.setPreviewCallback(this);

        mCamera.startPreview();

        mTextureView.setAlpha(1.0f);
        mTextureView.setRotation(90.0f);


        pixels = new int[pwidth*pheight];



    }

    @Override
    public void onSurfaceTextureSizeChanged(SurfaceTexture surface, int width, int height) {
        // Ignored, the Camera does all the work for us
    }

    @Override
    public boolean onSurfaceTextureDestroyed(SurfaceTexture surface) {
        mCamera.stopPreview();
        mCamera.release();
        mCamera.setPreviewCallback(null);

        return true;
    }

    @Override
    public void onSurfaceTextureUpdated(SurfaceTexture surface) {
        // Update your view here!
        mCamera.setPreviewCallback(this);


    }

    @Override
    public void onPreviewFrame(byte[] data, Camera camera) {

        data = rv.convertYUV2RGB(MainActivity.this,data,pwidth,pheight);

     //  sendbytearraytondkforopencvprocessing(data);








    }

    static {
        Native.register(MainActivity.class, "native-lib");
    }
    public native void toGrayScale(int pixels[], int len);




}

Renderscript.java

 public byte[] convertYUV2RGB(Context c,byte[] YUVArray, int H, int W){//W: 1280, H: 720

       RenderScript rs = RenderScript.create(c);
        ScriptIntrinsicYuvToRGB yuvToRgbIntrinsic = ScriptIntrinsicYuvToRGB.create(rs, Element.U8_4(rs));

        Type.Builder yuvType = new Type.Builder(rs, Element.U8(rs)).setX(YUVArray.length);

        Allocation in = Allocation.createTyped(rs, yuvType.create(), Allocation.USAGE_SCRIPT);

        Type.Builder rgbaType = new Type.Builder(rs, Element.RGBA_8888(rs)).setX(W).setY(H);

        Allocation out = Allocation.createTyped(rs, rgbaType.create(), Allocation.USAGE_SCRIPT);
        in.copyFrom(YUVArray);
        byte[] RGBOut = new byte[W * H * 4];


        yuvToRgbIntrinsic.setInput(in);
        yuvToRgbIntrinsic.forEach(out);

        out.copyTo(RGBOut);


        return RGBOut;
    }

1 个答案:

答案 0 :(得分:0)

  

我相信无法更新相机预览。

是的。

  

我需要在我的textureview顶部放置某种叠加层视图,然后将更新的字节数组呈现到叠加层视图的表面。

您可以在OpenCV中看到类似的方法。他们的Android演示显示RGB预览。他们不使用renderscript,但这没关系。无论如何,帧频(取决于CPU功率和相机分辨率)通常不够平滑。

要获得更好的性能,您可以使用纹理视图,或者更确切地说,可以通过直接在EGL纹理上设置像素(glTexSubImage2D)来使用GPU。参见example

但是,如果这样做,实际上会浪费大量时间在CPU和GPU之间传递数据:一次使用renderscript,然后再次设置纹理。相反,您可以使用OpenGL直接设置YUV像素,然后employ a shader to convert them to RGB for display