0

我有一个任务 - 缩小从相机获得的图像。我需要它来对较小版本的图像进行繁重的操作，这将帮助我节省一些处理能力。

我决定使用vImage_Bufferfrom Accelerate。这是我的代码，几乎没有评论，只是为了清楚地理解那里有什么：

        guard let imgBuffer = CMSampleBufferGetImageBuffer(buffer) else {
            return
        }


        CVPixelBufferLockBaseAddress(imgBuffer, CVPixelBufferLockFlags(rawValue: 0))

        // create vImage_Buffer out of CVImageBuffer
        var inBuff: vImage_Buffer = vImage_Buffer()
        inBuff.width = UInt(CVPixelBufferGetWidth(imgBuffer))
        inBuff.height = UInt(CVPixelBufferGetHeight(imgBuffer))
        inBuff.rowBytes = CVPixelBufferGetBytesPerRow(imgBuffer)
        inBuff.data = CVPixelBufferGetBaseAddress(imgBuffer)

        // bring down the size at half
        let new_width: UInt = inBuff.width/2
        let new_height: UInt = inBuff.height/2

        // create output buffer where scaled image is supposed to be
        var outBuff: vImage_Buffer = vImage_Buffer()
        outBuff.data = UnsafeMutableRawPointer.allocate(byteCount: Int(new_width * new_height * 4), alignment: MemoryLayout<UInt>.size)
        outBuff.width = new_width
        outBuff.height = new_height
        outBuff.rowBytes = Int(new_width * 4)

        // perform scale
        let err = vImageScale_ARGB8888(&inBuff, &outBuff, nil, 0)
        if err != kvImageNoError {
            print("Wrong!")
        }

        // I guess I need to unlock buffer at this point, right?
        CVPixelBufferUnlockBaseAddress(imgBuffer, CVPixelBufferLockFlags(rawValue: 0))

        // create CVImageBuffer
        let options = [kCVPixelBufferCGImageCompatibilityKey: true,
                       kCVPixelBufferCGBitmapContextCompatibilityKey: true,
                       kCVPixelBufferWidthKey: new_width,
                       kCVPixelBufferHeightKey: new_height] as CFDictionary

        var newPixelBuffer: CVImageBuffer?
        let status = CVPixelBufferCreateWithBytes(kCFAllocatorDefault,
                                                  Int(new_width), Int(new_height),
                                                  kCVPixelFormatType_32BGRA, &outBuff, Int(new_width * 4),
                                                  nil, nil, options, &newPixelBuffer)

        if status == kCVReturnError {
            print("Wrong again!")
        }

        // create CIImage from CVImageBuffer and UIImage from CIImage just to see how scale went
        let ciImg = CIImage(cvImageBuffer: newPixelBuffer!)
        let img = UIImage(ciImage: ciImg)
        delegate?.testSmallImage(img)

似乎所有操作都在没有任何错误的情况下执行，我想检查缩放是如何进行的，所以我试图创建新UIImage的缩放缓冲区。但是当我尝试用 显示图像时UIImageView，出现EXC_BAD_ACCESS错误。当我尝试保存时UIImage，一切都没有错误，但目录中没有文件出现Documents。你能指出我在这里到底做错了什么吗？谢谢！

4

1 回答 1

0

这是 Swift 上调整大小的代码片段CMSampleBuffer：

    private func scale(_ sampleBuffer: CMSampleBuffer) -> CVImageBuffer?
    {
        guard let imgBuffer = CMSampleBufferGetImageBuffer(sampleBuffer) else {
            return nil
        }

        CVPixelBufferLockBaseAddress(imgBuffer, CVPixelBufferLockFlags(rawValue: 0))

        // create vImage_Buffer out of CVImageBuffer
        var inBuff: vImage_Buffer = vImage_Buffer()
        inBuff.width = UInt(CVPixelBufferGetWidth(imgBuffer))
        inBuff.height = UInt(CVPixelBufferGetHeight(imgBuffer))
        inBuff.rowBytes = CVPixelBufferGetBytesPerRow(imgBuffer)
        inBuff.data = CVPixelBufferGetBaseAddress(imgBuffer)

        // perform scale
        var err = vImageScale_ARGB8888(&inBuff, &scaleBuffer, nil, 0)
        if err != kvImageNoError {
            print("Can't scale a buffer")
            return nil
        }
        CVPixelBufferUnlockBaseAddress(imgBuffer, CVPixelBufferLockFlags(rawValue: 0))

        var newBuffer: CVPixelBuffer?
        let attributes : [NSObject:AnyObject] = [
            kCVPixelBufferCGImageCompatibilityKey : true as AnyObject,
            kCVPixelBufferCGBitmapContextCompatibilityKey : true as AnyObject
        ]

        let status = CVPixelBufferCreateWithBytes(kCFAllocatorDefault,
                                                  Int(scaleBuffer.width), Int(scaleBuffer.height),
                                                  kCVPixelFormatType_32BGRA, scaleBuffer.data,
                                                  Int(scaleBuffer.width) * 4,
                                                  nil, nil,
                                                  attributes as CFDictionary?, &newBuffer)

        guard status == kCVReturnSuccess, let b = newBuffer else {
            print("Can't create new CVPixelBuffer")
            return nil
        }

        return b
    }

scaleBuffer这是一个在规模操作中充当目标的定义。我不需要每个比例都创建它，所以我只做一次：

        scaleBuffer.data = UnsafeMutableRawPointer.allocate(byteCount: Int(new_width * new_height * 4), alignment: MemoryLayout<UInt>.size)
        scaleBuffer.width = vImagePixelCount(new_width)
        scaleBuffer.height = vImagePixelCount(new_height)
        scaleBuffer.rowBytes = Int(new_width * 4)
于 2019-09-03T11:57:46.720 回答