I have checkout out the latest Google Vision APIs from here:
https://github.com/googlesamples/android-vision
And I am running it on a LG G2 device with KitKat. The only change I have made is to the minSdkVerion in the Gradle file:
...
defaultConfig {
applicationId "com.google.android.gms.samples.vision.face.multitracker"
minSdkVersion 19
...
However it does not focus. How do I make it focus?
I modified the CameraSourcePreview (....) constructor to be as follows:
public CameraSourcePreview(Context context, AttributeSet attrs) {
super(context, attrs);
mContext = context;
mStartRequested = false;
mSurfaceAvailable = false;
mSurfaceView = new SurfaceView(context);
mSurfaceView.getHolder().addCallback(new SurfaceCallback());
addView(mSurfaceView);
mSurfaceView.setOnClickListener(new OnClickListener() {
@Override
public void onClick(View v) {
cameraFocus(mCameraSource, Camera.Parameters.FOCUS_MODE_CONTINUOUS_VIDEO);
}
});
}
private static boolean cameraFocus(@NonNull CameraSource cameraSource, @NonNull String focusMode) {
Field[] declaredFields = CameraSource.class.getDeclaredFields();
for (Field field : declaredFields) {
if (field.getType() == Camera.class) {
field.setAccessible(true);
try {
Camera camera = (Camera) field.get(cameraSource);
if (camera != null) {
Camera.Parameters params = camera.getParameters();
params.setFocusMode(focusMode);
camera.setParameters(params);
return true;
}
return false;
} catch (IllegalAccessException e) {
e.printStackTrace();
}
break;
}
}
return false;
}
The advice was given here: https://github.com/googlesamples/android-vision/issues/2
and the code reference was here: https://gist.github.com/Gericop/7de0b9fdd7a444e53b5a
I also had to modify the FaceTrackerFactory draw(Canvas ...) method:
@Override
public void draw(Canvas canvas) {
Face face = mFace;
if (face == null) {
return;
}
// Draws a circle at the position of the detected face, with the face's track id below.
float cx = translateX(face.getPosition().x + face.getWidth() / 2);
float cy = translateY(face.getPosition().y + face.getHeight() / 2);
canvas.drawCircle(cx, cy, FACE_POSITION_RADIUS, mFacePositionPaint);
canvas.drawText("id: " + getId(), cx + ID_X_OFFSET, cy + ID_Y_OFFSET, mIdPaint);
// Draws an oval around the face.
float xOffset = scaleX(face.getWidth() / 2.0f);
float yOffset = scaleY(face.getHeight() / 2.0f);
float left = cx - xOffset;
float top = cy - yOffset;
float right = cx + xOffset;
float bottom = cy + yOffset;
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.LOLLIPOP) {
canvas.drawOval(left, top, right, bottom, mBoxPaint);
} else {
canvas.drawCircle(cx, cy, Math.max(xOffset, yOffset), mBoxPaint);
}
}
An auto focus option is now available in the official API. See the setAutoFocusEnabled method here:
Also, we open sourced the CameraSource class, which has an auto focus method as well. This one allows you to set a specific focus mode as opposed to the "continuous video" mode that the official API defaults to:
This worked for me Using Google Play Services 8.4: 'com.google.android.gms:play-services:8.4.0'
cameraSource = new CameraSource.Builder(this, detector).setRequestedPreviewSize(640, 480).setAutoFocusEnabled(true).build();
来源:https://stackoverflow.com/questions/32051973/google-vision-api-samples-get-the-camerasource-to-focus