我有一个SurfaceView
,可以将Bitmap
作为背景绘制,另一个将用作叠加层。所以我决定使用可以用于两个位图的Matrix
进行所有转换,因为它(我认为)是不使用OpenGL的最快方法之一。
我已经能够实现平移和缩放,但我遇到了一些问题:
有没有办法使用Matrix实现这个目标还是有另一种解决方案?
代码如下(我在单独的线程中使用SurfaceHolder
确实锁定SurfaceView画布并调用其doDraw
方法):
public class MapSurfaceView extends SurfaceView implements SurfaceHolder.Callback {
public void doDraw(Canvas canvas) {
canvas.drawColor(Color.BLACK);
canvas.drawBitmap(mBitmap, mTransformationMatrix, mPaintAA);
}
@Override
public boolean onTouchEvent(MotionEvent event) {
switch (event.getAction() & MotionEvent.ACTION_MASK) {
case MotionEvent.ACTION_POINTER_DOWN: {
if (event.getPointerCount() == 2) {
mOriginalDistance = MathUtils.distanceBetween(event.getX(0), event.getX(1), event.getY(0), event.getY(1));
mScreenMidpoint = MathUtils.midpoint(event.getX(0), event.getX(1), event.getY(0), event.getY(1));
mImageMidpoint = MathUtils.midpoint((mXPosition+event.getX(0))/mScale, (mXPosition+event.getX(1))/mScale, (mYPosition+event.getY(0))/mScale, (mYPosition+event.getY(1))/mScale);
mOriginalScale = mScale;
}
}
case MotionEvent.ACTION_DOWN: {
mOriginalTouchPoint = new Point((int)event.getX(), (int)event.getY());
mOriginalPosition = new Point(mXPosition, mYPosition);
break;
}
case MotionEvent.ACTION_MOVE: {
if (event.getPointerCount() == 2) {
final double currentDistance = MathUtils.distanceBetween(event.getX(0), event.getX(1), event.getY(0), event.getY(1));
if (mIsZooming || currentDistance - mOriginalDistance > mPinchToZoomTolerance || mOriginalDistance - currentDistance > mPinchToZoomTolerance) {
final float distanceRatio = (float) (currentDistance / mOriginalDistance);
float tempZoom = mOriginalScale * distanceRatio;
mScale = Math.min(10, Math.max(Math.min((float)getHeight()/(float)mBitmap.getHeight(), (float)getWidth()/(float)mBitmap.getWidth()), tempZoom));
mScale = (float) MathUtils.roundToDecimals(mScale, 1);
mIsZooming = true;
mTransformationMatrix = new Matrix();
mTransformationMatrix.setScale(mScale, mScale);//, mImageMidpoint.x, mImageMidpoint.y);
} else {
System.out.println("Dragging");
mIsZooming = false;
final int deltaX = (int) ((int) (mOriginalTouchPoint.x - event.getX()));
final int deltaY = (int) ((int) (mOriginalTouchPoint.y - event.getY()));
mXPosition = mOriginalPosition.x + deltaX;
mYPosition = mOriginalPosition.y + deltaY;
validatePositions();
mTransformationMatrix = new Matrix();
mTransformationMatrix.setScale(mScale, mScale);
mTransformationMatrix.postTranslate(-mXPosition, -mYPosition);
}
}
break;
}
case MotionEvent.ACTION_UP:
case MotionEvent.ACTION_POINTER_UP: {
mIsZooming = false;
validatePositions();
mTransformationMatrix = new Matrix();
mTransformationMatrix.setScale(mScale, mScale);
mTransformationMatrix.postTranslate(-mXPosition, -mYPosition);
}
}
return true;
}
private void validatePositions() {
// Lower right corner
mXPosition = Math.min(mXPosition, (int)((mBitmap.getWidth() * mScale)-getWidth()));
mYPosition = Math.min(mYPosition, (int)((mBitmap.getHeight() * mScale)-getHeight()));
// Upper left corner
mXPosition = Math.max(mXPosition, 0);
mYPosition = Math.max(mYPosition, 0);
// Image smaller than the container, should center it
if (mBitmap.getWidth() * mScale <= getWidth()) {
mXPosition = (int) -((getWidth() - (mBitmap.getWidth() * mScale))/2);
}
if (mBitmap.getHeight() * mScale <= getHeight()) {
mYPosition = (int) -((getHeight() - (mBitmap.getHeight() * mScale))/2);
}
}
}
答案 0 :(得分:2)
不是每次使用新的Matrix()重置转换矩阵,而是尝试使用post *()更新它。这样,您只进行相对于屏幕的操作。用术语来思考更容易:“在屏幕上缩放到这一点”。
现在有些代码了。在缩放部分计算了mScale:
...
mScale = (float) MathUtils.roundToDecimals(mScale, 1);
float ratio = mScale / mOriginalScale;
mTransformationMatrix.postScale(ratio, ratio, mScreenMidpoint.x, mScreenMidpoint.y);
在每个缩放触摸事件上重新计算mScreenMidpoint可能更好。这将允许用户在缩放时稍微改变焦点。对我来说,比前两个手指触摸后焦点冻结更自然。
在拖动过程中,您使用deltaX和deltaY而不是绝对点进行翻译:
mTransformationMatrix.postTranslate(-deltaX, -deltaY);
当然,您现在必须将validatePositions()方法更改为:
我将描述第二种方法,因为它更灵活,也允许验证缩放。
我们计算屏幕外的图像数量,然后使用这些值移动它:
void validate() {
mTransformationMatrix.mapRect(new RectF(0, 0, mBitmap.getWidth(), mBitmap.getHeight()));
float height = rect.height();
float width = rect.width();
float deltaX = 0, deltaY = 0;
// Vertical delta
if (height < mScreenHeight) {
deltaY = (mScreenHeight - height) / 2 - rect.top;
} else if (rect.top > 0) {
deltaY = -rect.top;
} else if (rect.bottom < mScreenHeight) {
deltaY = mScreenHeight - rect.bottom;
}
// Horziontal delta
if (width < mScreenWidth) {
deltaX = (mScreenWidth - width) / 2 - rect.left;
} else if (rect.left > 0) {
deltaX = -rect.left;
} else if (rect.right < mScreenWidth) {
deltaX = mScreenWidth - rect.right;
}
mTransformationMatrix.postTranslate(deltaX, deltaY)
}