amaldona / documentdetection

DocumentDetection

IMPORTANT: Download OpenCV framework from: https://github.com/opencv/opencv/releases and copy opencv2.framework to DocumentDetection/DocumentDetectionSDK/DocumentDetectionSDK/libs

Framework files

• VideoPreviewViewController: Parent class of DocumentDetectionViewController. This view controller only displays a layer with the video preview, and constantly updates a snapshot.
• DocumentDetectionViewController: subclass of VideoPreviewViewController. Displays another layer with the detection box over the video preview layer, and interacts with the DocumentDetectionModel to obtain the corners.
• DocumentDetectionModel: Model class that interacts with the OpenCV wrapper, detects the corners, crops and transforms the snapshot, and stores the result.
• CGPoint+Helper: Helper class with some point operations like transforming a point from one coordinate system to another, sort detected points, etc.
• CIImage+Helper: Helper class with cropping and perspective correction methods.
• OpenCVWrapper: C++ wrapper class.
• DocumentDetection: Entry point of the framework. Only has one static method that returns the transformed image.
• Tests with a Host Application target.

The workspace includes a barebone sample project, with a button that calls the framework entry point method (DocumentDetection class) and an imageView to display the result:

DocumentDetection.scanDocument { image, error in

	// Show image result
	if let image = image {
	    self.image.image = image
	}

	// Show error message
	if let error = error {
	    let message = switch error {
	        case .ImageNotFound:
	            "Image not found"
	        case .CropFailed:
	            "Error cropping the image"
	        case .TransformFailed:
	            "Error applying perspective correction to the image"
	        default:
	            "Unknown error"
	    }
	    print(message)
	}

}