Question

我使用Vision进行文本检测，但这显示了所有屏幕上的检测，有没有办法让它只检测某个区域，例如我在屏幕中间绘制的矩形。

因此，当整个屏幕是一个摄像机，但如果文本在中心进入一个矩形，它会围绕它绘制框。

以下是我的文字检测代码和cameraPreviewLayer。

如果可能，我需要2个相机预览图层吗？

func setupPreviewLayer() {
    cameraPreviewlayer = AVCaptureVideoPreviewLayer(session: captureSession)
    cameraPreviewlayer?.videoGravity = AVLayerVideoGravity.resizeAspectFill
    cameraPreviewlayer?.connection?.videoOrientation = AVCaptureVideoOrientation.portrait
    cameraPreviewlayer?.frame = self.view.frame
    self.view.layer.insertSublayer(cameraPreviewlayer!, at: 0)
}

func startTextDetection() {
    let textRequest = VNDetectTextRectanglesRequest(completionHandler: self.detectTextHandler)
    textRequest.reportCharacterBoxes = true
    self.requests = [textRequest]
}

func detectTextHandler(request: VNRequest, error: Error?) {
    guard let observations = request.results else {
        print("no result")
        return
    }

    let result = observations.map({$0 as? VNTextObservation})

    DispatchQueue.main.async() {
        self.cameraPreviewlayer?.sublayers?.removeSubrange(1...)
        for region in result {
            guard let rg = region else {
                continue
            }

            self.highlightWord(box: rg)

            if let boxes = region?.characterBoxes {
                for characterBox in boxes {
                    self.highlightLetters(box: characterBox)
                }
            }
        }
    }
}

func highlightWord(box: VNTextObservation) {
    guard let boxes = box.characterBoxes else {
        return
    }

    var maxX: CGFloat = 9999.0
    var minX: CGFloat = 0.0
    var maxY: CGFloat = 9999.0
    var minY: CGFloat = 0.0

    for char in boxes {
        if char.bottomLeft.x < maxX {
            maxX = char.bottomLeft.x
        }
        if char.bottomRight.x > minX {
            minX = char.bottomRight.x
        }
        if char.bottomRight.y < maxY {
            maxY = char.bottomRight.y
        }
        if char.topRight.y > minY {
            minY = char.topRight.y
        }
    }

    let xCord = maxX * (cameraPreviewlayer?.frame.size.width)!
    let yCord = (1 - minY) * (cameraPreviewlayer?.frame.size.height)!
    let width = (minX - maxX) * (cameraPreviewlayer?.frame.size.width)!
    let height = (minY - maxY) * (cameraPreviewlayer?.frame.size.height)!

    let outline = CALayer()
    outline.frame = CGRect(x: xCord, y: yCord, width: width, height: height)
    outline.borderWidth = 2.0
    outline.borderColor = UIColor.red.cgColor

    cameraPreviewlayer?.addSublayer(outline)
}

func highlightLetters(box: VNRectangleObservation) {
    let xCord = box.topLeft.x * (cameraPreviewlayer?.frame.size.width)!
    let yCord = (1 - box.topLeft.y) * (cameraPreviewlayer?.frame.size.height)!
    let width = (box.topRight.x - box.bottomLeft.x) * (cameraPreviewlayer?.frame.size.width)!
    let height = (box.topLeft.y - box.bottomLeft.y) * (cameraPreviewlayer?.frame.size.height)!

    let outline = CALayer()
    outline.frame = CGRect(x: xCord, y: yCord, width: width, height: height)
    outline.borderWidth = 1.0
    outline.borderColor = UIColor.blue.cgColor

    cameraPreviewlayer?.addSublayer(outline)
}

Answer 1

我通过添加来解决这个问题。

let wordRect = CGRect(x: xCord, y: yCord, width: width, height: height)
    guard regionOfInterest.contains(wordRect.origin) else { return } // only draw a box if the orgin of the word box is within the regionOfInterest

之前的

 let outline = CALayer()
outline.frame = CGRect(x: xCord, y: yCord, width: width, height: height)
outline.borderWidth = 1.0
outline.borderColor = UIColor.blue.cgColor

cameraPreviewlayer?.addSublayer(outline)

这样它只会运行矩形中单词的代码我将单词rect设置为uiview大小的框架，我只想显示文本。

检测带有矩形的文本而不是整个屏幕的视觉，iOS，Swift

1 个答案: