最近做了很多客制化的Touch需求，很多情况都可以直接通过GestureDetector来完成，这里挑选三指截屏简述原理。先来说说主要思路，在view的onInterceptTouchEvent方法中，处理pointerCount==3时的ACTION_MOVE事件，在这里可以判断手指滑动的速度，距离等，当满足我们需要的条件时，便可进行操作(这里的截屏是通过sendBroadcast(intent)后在PhoneWindowManager中接受广播进行截屏操作)。

在重写相关view的Touch操作时，先获取到当前的屏幕参数：

DisplayMetrics metrics = getContext().getResources().getDisplayMetrics();//获取手机屏幕参数

mHeightPixels = metrics.heightPixels;//获取手机屏幕高度

重写onInterceptTouchEvent，利用Point存储手指移动前后X、Y轴的位置，可以在event.getPointerCount()满足需求时进行判断(不同屏幕支持多点触控的count也不相同，自己测试得到的最大值为5)，同时定义了一个mMode变量，可以在case MotionEvent.ACTION_POINTER_DOWN中为其赋值，设置当前的mode类型，具体代码如下：

public boolean onInterceptTouchEvent(MotionEvent event) {

if(true) {

int pointerCount = event.getPointerCount();

switch (event.getAction() & MotionEvent.ACTION_MASK) {

case MotionEvent.ACTION_DOWN:

mEnterTime = System.currentTimeMillis();

mMode = MODE_NONE;

CURRENT_SOUND_STEP = 15;//mAM.getStreamVolume(AudioManager.STREAM_SYSTEM);

initP0.x = preP0.x = (int)event.getX(0);

initP0.y = preP0.y = (int)event.getY(0);

break;

case MotionEvent.ACTION_MOVE:

//for when pointerCount ==3 but not go into ACTION_POINTER_DOWN

if(mMode == MODE_NONE || mMode == MODE_INTERUPT) {

break;

} else if((mMode == MODE_2) && (pointerCount == 2)) {

Point current0 = new Point((int)event.getX(0), (int)event.getY(0));

Point current1 = new Point((int)event.getX(1), (int)event.getY(1));

int step = (current0.y - preP0.y) / SOUND_STEP;

if(step != 0) {

int current = CURRENT_SOUND_STEP + (initP0.y - current0.y) / SOUND_STEP;

current = current < 0 ? 0 : current;

current = current > MAX_SOUND_STEP ? MAX_SOUND_STEP : current;

Log.i(TAG, "current:" + current);

//mAM.setMode(AudioManager.MODE_NORMAL);

mAM.setStreamVolume(AudioManager.STREAM_SYSTEM, current, AudioManager.FLAG_PLAY_SOUND);

preP0 = current0;

preP1 = current1;

} else if((mMode == MODE_3) && (pointerCount == 3)) {

Point current0 = new Point((int)event.getX(0), (int)event.getY(0));

Point current1 = new Point((int)event.getX(1), (int)event.getY(1));

Point current2 = new Point((int)event.getX(2), (int)event.getY(2));

if(Math.abs(current0.y - initP0.y) > (mHeightPixels / 10) &&

Math.abs(current1.y - initP1.y) > (mHeightPixels / 10) &&

Math.abs(current2.y - initP2.y) > (mHeightPixels / 10)) {

Intent intent = new Intent("intent.action.SendKey");

intent.putExtra("KeyCode", KeyEvent.KEYCODE_SYSRQ);

getContext().sendBroadcast(intent);

Log.e(TAG, "takeScreenshot!");

mMode = MODE_INTERUPT;

}

return true;//该touch事件终止向子view传递

}

break;

//有一个非主要的手指按下了

case MotionEvent.ACTION_POINTER_DOWN:

if(mMode == MODE_INTERUPT) {

break;

}

if(pointerCount == 2) {

mMode = MODE_2;

initP1.x = preP1.x = (int)event.getX(1);

initP1.y = preP1.y = (int)event.getY(1);

if(Math.abs(initP0.y - initP1.y) > (SOUND_STEP * 2)) {

mMode = MODE_INTERUPT;

Log.i(TAG, "ACTION_POINTER_DOWN p0 and p1 are too far.");

break;

}

Log.i(TAG, "ACTION_POINTER_DOWN MODE_2.");

} else if(pointerCount == 3) {

mMode = MODE_3;

initP1.x = preP1.x = (int)event.getX(1);

initP1.y = preP1.y = (int)event.getY(1);

initP2.x = preP2.x = (int)event.getX(2);

initP2.y = preP2.y = (int)event.getY(2);

Log.i(TAG, "ACTION_POINTER_DOWN MODE_3.");

}

if((System.currentTimeMillis() - mEnterTime) > 500) {

mMode = MODE_INTERUPT;

Log.i(TAG, "ACTION_POINTER_DOWN too slowly.");

}

break;

//有一个非主要的手指抬起了

case MotionEvent.ACTION_POINTER_UP:

mMode = MODE_INTERUPT;

break;

case MotionEvent.ACTION_UP:

mMode = MODE_NONE;

break;

}

}

}

具体的滑动距离，速度，方向等都可已自己根据需求修改。

接收广播时，registerReceiver与unregisterReceiver即可，注意在进行截屏前判断下当前的开关值

BroadcastReceiver mKeyReceiver = new BroadcastReceiver() {

@Override

public void onReceive(Context context, Intent intent) {

String action = intent.getAction();

boolean m3PTouchScreenshot;

if ("intent.action.SendKey".equals(action)) {

Bundle bundle = intent.getExtras();

int keycode = bundle.getInt("KeyCode");

Log.i(TAG, "mKeyReceiver:" + keycode);

switch(keycode) {

case KeyEvent.KEYCODE_SYSRQ:

boolean touchScreenshot = SystemProperties.getBoolean("3p_screenshot_off", false);

m3PTouchScreenshot = Settings.Global.getInt(context.getContentResolver(), "3p_touch_screenshot", (!touchScreenshot ? 1 : 0)) == 1;

Log.i(TAG, "m3PTouchScreenshot:" + m3PTouchScreenshot);

if(m3PTouchScreenshot)

takeScreenshot();

break;

}

}

}

};

takeScreenshot()就是执行的截屏操作啦～