触屏设备上的多点触碰检测C++代码实现

转自：http://aigo.iteye.com/blog/2272698

代码还是参考自Epic官方的塔防项目：StrategyGame

看了下C++的API，现成的API中貌似只支持单点触碰检测，用法如下：
注册：

// support touch devices

    InputComponent->BindTouch(EInputEvent::IE_Pressed, this, &ATD_MobilePlayerController::MoveToTouchLocation);

    InputComponent->BindTouch(EInputEvent::IE_Repeat, this, &ATD_MobilePlayerController::MoveToTouchLocation);  

触发回调：

void ATD_MobilePlayerController::MoveToTouchLocation(const ETouchIndex::Type FingerIndex, const FVector Location)

{

    FVector2D ScreenSpaceLocation(Location);

    // Trace to see what is under the touch location

    FHitResult HitResult;

    GetHitResultAtScreenPosition(ScreenSpaceLocation, CurrentClickTraceChannel, true, HitResult);

    if (HitResult.bBlockingHit)

    {

        // We hit something, move there

        SetNewMoveDestination(HitResult.ImpactPoint);

    }

}  

如果要实现多点触碰检测，比如实现两个手指捏动来缩放屏幕，StrategyGame项目自己实现了一套算法来实现多点屏幕检测，具体做法如下：
StrategyPlayerController.h

1，先重写StrategyPlayerController的ProcessPlayerInput()和SetupInputComponent()函数

protected:

    /** update input detection */

    virtual void ProcessPlayerInput(const float DeltaTime, const bool bGamePaused) override;

    virtual void SetupInputComponent() override;  

在SetupInputComponent()函数中绑定事件（这个事件机制也是自己写的），其中BIND_1P_ACTION和BIND_2P_ACTION是自己定义的宏：

ProcessPlayerInput()来检测触碰点变化，这是一个连续执行的函数，大概逻辑是：每次执行时都会把当前的屏幕触碰信息和上次的触碰信息做对比，检测是单点触碰，还是按住不放，还是多点触碰或多点按住不放等等，具体逻辑在InputHandler->UpdateDetection(DeltaTime);中。

void AStrategyPlayerController::ProcessPlayerInput(const float DeltaTime, const bool bGamePaused)

{

    if (!bGamePaused && PlayerInput && InputHandler && !bIgnoreInput)

    {

        InputHandler->UpdateDetection(DeltaTime);

    }

    Super::ProcessPlayerInput(DeltaTime, bGamePaused);

    if (!bIgnoreInput )

    {

        const ULocalPlayer* LocalPlayer = Cast<ULocalPlayer>(Player);

        AStrategySpectatorPawn* StrategyPawn = GetStrategySpectatorPawn();

        if(( StrategyPawn != NULL ) && ( LocalPlayer != NULL ))

        {

            // Create the bounds for the minimap so we can add it as a 'no scroll' zone.

            AStrategyHUD* const HUD = Cast<AStrategyHUD>(GetHUD());

            AStrategyGameState const* const MyGameState = GetWorld()->GetGameState<AStrategyGameState>();

            if( (MyGameState != NULL ) && ( MyGameState->MiniMapCamera.IsValid() == true ) )

            {

                if( LocalPlayer->ViewportClient != NULL )

                {

                    const FIntPoint ViewportSize = LocalPlayer->ViewportClient->Viewport->GetSizeXY();

                    const uint32 ViewTop = FMath::TruncToInt(LocalPlayer->Origin.Y * ViewportSize.Y);

                    const uint32 ViewBottom = ViewTop + FMath::TruncToInt(LocalPlayer->Size.Y * ViewportSize.Y);

                    FVector TopLeft( HUD->MiniMapMargin, ViewBottom - HUD->MiniMapMargin - MyGameState->MiniMapCamera->MiniMapHeight,  );

                    FVector BottomRight( (int32)MyGameState->MiniMapCamera->MiniMapWidth, MyGameState->MiniMapCamera->MiniMapHeight,  );

                    FBox MiniMapBounds( TopLeft, TopLeft + BottomRight );

                    StrategyPawn->GetStrategyCameraComponent()->AddNoScrollZone( MiniMapBounds );

                    StrategyPawn->GetStrategyCameraComponent()->UpdateCameraMovement( this );

                }

            }

        }

    }

}  

StrategyInput.cpp