boofcv.factory.fiducial.ConfigMicroQrCode Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of boofcv-recognition Show documentation
BoofCV is an open source Java library for real-time computer vision and robotics applications.
The newest version!
/*
 * Copyright (c) 2023, 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.factory.fiducial;

import boofcv.abst.shapes.polyline.ConfigPolylineSplitMerge;
import boofcv.alg.fiducial.qrcode.EciEncoding;
import boofcv.factory.filter.binary.ConfigThreshold;
import boofcv.factory.filter.binary.ConfigThresholdLocalOtsu;
import boofcv.factory.filter.binary.ThresholdType;
import boofcv.factory.shape.ConfigPolygonDetector;
import boofcv.struct.ConfigLength;
import boofcv.struct.Configuration;
import boofcv.struct.ConnectRule;
import org.jetbrains.annotations.Nullable;

/**
 * Configuration for {@link boofcv.abst.fiducial.MicroQrCodePreciseDetector}
 *
 * @author Peter Abeles
 */
public class ConfigMicroQrCode implements Configuration {
	/** Specifies how images are thresholded and converted into a binary format */
	public ConfigThreshold threshold;

	/** Configuration for polygon detector that's used to find position patterns */
	public ConfigPolygonDetector polygon = new ConfigPolygonDetector();

	/**
	 * If not null, then when decoding BYTE mode data it will always use this encoding. This can be desirable
	 * if the automatic encoding detection is making a mistake or if you know the data is binary. For binary
	 * data you should set this to {@link EciEncoding#BINARY}.
	 */
	public @Nullable String forceEncoding = null;

	/**
	 * Fore BYTE mode, if the auto encoding detection decides it's not UTF-8 then it will use this encoding.
	 * Depending on which QR code standard you are following (few people follow either) it should be
	 * {@link EciEncoding#ISO8859_1} or {@link EciEncoding#JIS}.
	 */
	public String defaultEncoding = EciEncoding.ISO8859_1;

	/**
	 * If true it will consider QR codes which have been incorrectly encoded with transposed bits. Set to false if
	 * you know your markers are standard compliant and want a modest speed boost.
	 */
	public boolean considerTransposed = true;

	/**
	 * This turns off the check to ensure padding bytes have the expected pattern. This was added due to a bug
	 * in a popular encoder where for messages of a certain length it would be off by one. Since no data
	 * is encoded in the padding and bugs are so common in encoders, by default we will ignore the padding.
	 */
	public boolean ignorePaddingBytes = true;

	{
		// 40% slower but better at detecting fiducials by a few percentage points
//		ConfigThreshold configThreshold = ConfigThreshold.local(ThresholdType.LOCAL_MEAN,15);
//		configThreshold.scale = 1.00;

		// fast but does a bad job detecting fiducials that are up close
		ConfigThresholdLocalOtsu configThreshold = ConfigThreshold.local(ThresholdType.BLOCK_OTSU, 40);
		configThreshold.useOtsu2 = true;
		// 0.95 makes it better some times but worse overall
		configThreshold.scale = 1.0;
		// this will hurt small distant targets but allows up close to work
		configThreshold.thresholdFromLocalBlocks = true;
		configThreshold.tuning = 4;

		threshold = configThreshold;

		polygon.detector.contourRule = ConnectRule.EIGHT;
		polygon.detector.clockwise = false;
		((ConfigPolylineSplitMerge)polygon.detector.contourToPoly).maxSideError = ConfigLength.relative(0.12, 3);
		((ConfigPolylineSplitMerge)polygon.detector.contourToPoly).cornerScorePenalty = 0.4;
		((ConfigPolylineSplitMerge)polygon.detector.contourToPoly).minimumSideLength = 2;
		// 28 pixels = 7 by 7 square viewed head on. Each cell is then 1 pixel. Any slight skew results in
		// aliasing and will most likely not be read well.
		polygon.detector.minimumContour = ConfigLength.fixed(40);
		// can handle much darker images. No measurable decrease in speed
		polygon.detector.minimumEdgeIntensity = 3;
		polygon.minimumRefineEdgeIntensity = 6;
		// TODO This needs to be reduced for smaller shapes, but should be larger to better handle blur? Experiment
		polygon.detector.tangentEdgeIntensity = 1.5;
	}

	/**
	 * Default configuration for a QR Code detector which is optimized for speed
	 */
	public static ConfigMicroQrCode fast() {
		// A global threshold is faster than any local algorithm
		// plus it will generate a simpler set of internal contours speeding up that process
		ConfigMicroQrCode config = new ConfigMicroQrCode();
		config.threshold = ConfigThreshold.global(ThresholdType.GLOBAL_OTSU);
		return config;
	}

	public ConfigMicroQrCode setTo( ConfigMicroQrCode src ) {
		this.threshold.setTo(src.threshold);
		this.polygon.setTo(src.polygon);
		this.forceEncoding = src.forceEncoding;
		this.defaultEncoding = src.defaultEncoding;
		this.considerTransposed = src.considerTransposed;
		this.ignorePaddingBytes = src.ignorePaddingBytes;
		return this;
	}

	@Override
	public void checkValidity() {
		threshold.checkValidity();
		polygon.checkValidity();
	}
}