GPUImage2 安装和简单使用

前言：发现网上GPUImage2的资源比较少，我最近又在使用这个库，所以我就把我使用的情况总结下吧。需要的人自取哦！

首先感谢@Willie文章提供参考

安装

1.pod安装

传送门
这个是英文版的。内容如下

下载GPUImage2.podspec文件下载地址(路径与Podfle同层级)
Podfile中添加 pod 'GPUImage2', :podspec => './GPUImage2.podspec'
pod install
import GPUImage2

2.直接导入库
原库下载地址

同@Willie文章导入 不同在于：不导入Source文件夹，其他一样

如果需要更改库的话，建议直接导入库的方式使用

简单使用

1.创建滤镜视频

do {

    camera = try Camera(sessionPreset: .vga640x480,
                        cameraDevice: nil,
                        location: .frontFacing,
                        captureAsYUV: true)

    
} catch {
    
    print(error)
    return
}
//磨皮 - 美白
let basicOperation = BilateralBlur()
basicOperation.distanceNormalizationFactor = 7
        
let brightnessAdjustment = BrightnessAdjustment()
brightnessAdjustment.brightness = 0.11

//显示
renderView = RenderView(frame: view.bounds)
view.addSubview(renderView)

camera --> brightnessAdjustment --> basicOperation --> renderView
basicOperation --> self.videoOutput

2.获取滤镜后的视频流

// 全局变量
videoOutput = RawDataOutput()
   
// 这里是修改源码，增加的一个方法（修改见下3）     
videoOutput.dataAvailableCallbackSize = { [weak self] (videoData, framebuffer, cmtime) in
    
    guard let `self` = self else { return }
    
    let numberOfBytesPerRow = framebuffer.size.width
    let getData = Data(bytes: videoData)
    
    getData.withUnsafeBytes { (u8Ptr: UnsafePointer<UInt8>) -> Void in
        let rawPtr = UnsafeMutableRawPointer(mutating: u8Ptr)
        var pixelBuffer : CVPixelBuffer?;
        _ = CVPixelBufferCreateWithBytes(kCFAllocatorDefault,
                                         Int(framebuffer.size.width),
                                         Int(framebuffer.size.height),
                                         kCVPixelFormatType_32BGRA,
                                         rawPtr,
                                         Int(numberOfBytesPerRow*4), nil, nil, nil,
                                         &pixelBuffer);
        if pixelBuffer != nil {
            
            // 获取到后，做点什么？
            
        }
    }
}

3.修改`RawDataOutput`

一、添加属性

public var dataAvailableCallbackSize:(([UInt8], _ frameBuffer: Framebuffer, _ cmtime: CMTime) -> ())?

二、在newFramebufferAvailable方法最后面调用

var time: CMTime = CMTime(seconds: 0, preferredTimescale: 1000)
if let timestamp = framebuffer.timingStyle.timestamp {
    time = CMTime(value: timestamp.value, timescale: timestamp.timescale)
}
dataAvailableCallbackSize?(data, framebuffer, time)

三、会出现蓝色阴影？？？

修改文件中 GL_RGBA为GL_BGRA

glReadPixels(0, 0, framebuffer.size.width, framebuffer.size.height, GLenum(GL_BGRA), GLenum(GL_UNSIGNED_BYTE), &data)

4.切换摄像头，修改 `Camera`

增加方法，外部直接调用即可

public func switchCamera() {
    
    self.location = self.location == .backFacing ? .frontFacing : .backFacing;

}

修改原属性 location

public var location:PhysicalCameraLocation {
    didSet {
        // TODO: Swap the camera locations, framebuffers as needed
        guard let newVideoInput = try? AVCaptureDeviceInput(device: location.device()!) else { return }
        
        self.captureSession.beginConfiguration()
        self.captureSession.removeInput(self.videoInput)
        self.captureSession.addInput(newVideoInput)
        self.captureSession.commitConfiguration()
        self.videoInput = newVideoInput
    }
}

5.视屏镜像

     var captureConnection: AVCaptureConnection!
    for connection in videoOutput.connections {
        for port in connection.inputPorts {
            if (port as AnyObject).mediaType == AVMediaType.video {
                captureConnection = connection
                captureConnection.isVideoMirrored = location == .frontFacing
            }
        }
    }
    
    if captureConnection.isVideoOrientationSupported {
        captureConnection.videoOrientation = .portraitUpsideDown
    }

如果这个文章帮到了你，一定给我Star、点击关注哦！

GPUImage2 安装和简单使用

安装

简单使用

1.创建滤镜视频

2.获取滤镜后的视频流

3.修改RawDataOutput

4.切换摄像头，修改 Camera

5.视屏镜像

推荐阅读更多精彩内容

3.修改`RawDataOutput`

4.切换摄像头，修改 `Camera`