boofcv.alg.transform.pyramid.PyramidDiscreteSampleBlur Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of boofcv-ip Show documentation
BoofCV is an open source Java library for real-time computer vision and robotics applications.
The newest version!
/*
 * Copyright (c) 2021, Peter Abeles. All Rights Reserved.
 *
 * This file is part of BoofCV (http://boofcv.org).
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package boofcv.alg.transform.pyramid;

import boofcv.abst.filter.convolve.ConvolveDown;
import boofcv.factory.filter.convolve.FactoryConvolveDown;
import boofcv.struct.border.BorderType;
import boofcv.struct.convolve.Kernel1D;
import boofcv.struct.image.ImageBase;
import boofcv.struct.image.ImageType;
import boofcv.struct.pyramid.ConfigDiscreteLevels;
import boofcv.struct.pyramid.ImagePyramid;
import boofcv.struct.pyramid.PyramidDiscrete;
import org.jetbrains.annotations.Nullable;

/**
 * 
 * Convolves a re-normalizable blur kernel across the image before down sampling. This is useful for creating
 * a Gaussian pyramid as well as other standard pyramids.
 * 
 *
 * 
 * NOTE: This pyramid cannot be configured such that blur will be applied to the input image. It can be done by
 * the user before the image is passed in.
 * 
 *
 * 
 * NOTE: This blur magnitude is constant for each level in the pyramid. In general it is desirable to
 * have it dependent on each level's scale factor.
 * 
 *
 * @author Peter Abeles
 */
@SuppressWarnings({"unchecked", "NullAway.Init"})
public class PyramidDiscreteSampleBlur> extends PyramidDiscrete {

	// stores the results from the first convolution
	private T temp;
	ConvolveDown horizontal;
	ConvolveDown vertical;

	// amount of blur applied to each layer
	double[] sigmas;
	private double sigma;

	/**
	 *
	 * @param kernel A blur kernel
	 * @param sigma The effective amount of Gaussian blur the kernel applies
	 * @param imageType Type of image processed
	 * @param saveOriginalReference If a reference to the full resolution image should be saved instead of copied.
	 *                              Set to false if you don't know what you are doing.
	 * @param configLayers Specifies how the levels are computed
	 */
	public PyramidDiscreteSampleBlur(Kernel1D kernel, double sigma, ImageType imageType,
									 boolean saveOriginalReference,
									 @Nullable ConfigDiscreteLevels configLayers)
	{
		super(imageType,saveOriginalReference,configLayers);

		horizontal = FactoryConvolveDown.convolve(kernel, BorderType.NORMALIZED, true, 1, imageType,imageType);
		vertical = FactoryConvolveDown.convolve(kernel, BorderType.NORMALIZED, false, 1, imageType,imageType);
		this.sigma = sigma;
	}

	private void lazyDeclareSigmas(double sigma) {
		if( this.sigmas != null )
			return;
		sigmas = new double[ super.levelScales.length ];
		sigmas[0] = 0;
		for( int i = 1; i < sigmas.length; i++ ) {
			// blur in previous layer
			double prev = sigmas[i-1];
			// the effective amount of blur applied to previous layer while being down sampled
			double applied = sigma*super.levelScales[i-1];
			// The amount of blur which has been applied to this layer
			sigmas[i] = Math.sqrt(prev*prev + applied*applied);
		}
	}

	protected PyramidDiscreteSampleBlur( PyramidDiscreteSampleBlur orig )
	{
		super(orig);

		this.horizontal = orig.horizontal;
		this.vertical = orig.vertical;
		this.sigma = orig.sigma;
	}

	@Override
	public void process(T input) {
		super.initialize(input.width,input.height);
		lazyDeclareSigmas(this.sigma);

		if( temp == null ) {
			// declare it to be the latest image that it might need to be, resize below
			temp = (T)input.createNew(1,1);
		}

		if (levelScales[0] == 1) {
			if (isSaveOriginalReference()) {
				setFirstLayer(input);
			} else {
				getLayer(0).setTo(input);
			}
		} else {
			int skip = levelScales[0];

			horizontal.setSkip(skip);
			vertical.setSkip(skip);

			temp.reshape(input.width/skip,input.height);
			horizontal.process(input,temp);
			vertical.process(temp,getLayer(0));
		}

		for (int index = 1; index < getNumLayers(); index++) {
			int skip = levelScales[index]/ levelScales[index-1];
			T prev = getLayer(index-1);
			temp.reshape(prev.width/skip,prev.height);

			horizontal.setSkip(skip);
			vertical.setSkip(skip);

			horizontal.process(prev,temp);
			vertical.process(temp,getLayer(index));
		}
	}

	/**
	 * There is no offset since a symmetric kernel is applied starting at pixel (0,0)
	 *
	 * @param layer Layer in the pyramid
	 * @return offset
	 */
	@Override
	public double getSampleOffset(int layer) {
		return 0;
	}

	@Override
	public double getSigma(int layer) {
		return sigmas[layer];
	}

	@Override
	public ImagePyramid copyStructure() {
		return new PyramidDiscreteSampleBlur<>(this);
	}
}