最好放在Util.java类中

public static float dpFromPx(final Context context, final float px) {
    return px / context.getResources().getDisplayMetrics().density;
}

public static float pxFromDp(final Context context, final float dp) {
    return dp * context.getResources().getDisplayMetrics().density;
}

更优雅的方法是使用kotlin的扩展函数

/**
 * Converts dp to pixel
 */
val Int.dpToPx: Int get() = (this * Resources.getSystem().displayMetrics.density).toInt()

/**
 * Converts pixel to dp
 */
val Int.pxToDp: Int get() = (this / Resources.getSystem().displayMetrics.density).toInt()

用法:

println("16 dp in pixel: ${16.dpToPx}")
println("16 px in dp: ${16.pxToDp}")

没有Context，优雅的静态方法:

public static int dpToPx(int dp)
{
    return (int) (dp * Resources.getSystem().getDisplayMetrics().density);
}

public static int pxToDp(int px)
{
    return (int) (px / Resources.getSystem().getDisplayMetrics().density);
}

Java代码:

// Converts 14 dip into its equivalent px
float dip = 14f;
Resources r = getResources();
float px = TypedValue.applyDimension(
    TypedValue.COMPLEX_UNIT_DIP,
    dip,
    r.getDisplayMetrics()
);

芬兰湾的科特林代码:

 val dip = 14f
 val r: Resources = resources
 val px = TypedValue.applyDimension(
     TypedValue.COMPLEX_UNIT_DIP,
     dip,
     r.displayMetrics
 )

芬兰湾的科特林扩展:

val Number.toPx get() = TypedValue.applyDimension(
  TypedValue.COMPLEX_UNIT_DIP,
  this.toFloat(),
  Resources.getSystem().displayMetrics)

上面有很多很棒的解决方案。然而，我发现最好的解决方案是谷歌的设计:

https://design.google.com/devices/

float scaleValue = getContext().getResources().getDisplayMetrics().density;
int pixels = (int) (dps * scaleValue + 0.5f);