Java tutorial
/* * To change this license header, choose License Headers in Project Properties. * To change this template file, choose Tools | Templates * and open the template in the editor. */ package pt.scanner.server.main; import com.google.common.base.Strings; import com.vividsolutions.jts.geom.*; import java.nio.*; import java.util.*; import java.util.stream.*; import org.apache.commons.math3.util.*; import org.bytedeco.javacpp.*; import org.bytedeco.javacpp.opencv_core.Mat; import org.bytedeco.javacpp.opencv_core.MatVector; import org.bytedeco.javacpp.opencv_core.Scalar; import org.bytedeco.javacpp.opencv_core.Size; import static org.bytedeco.javacpp.opencv_core.*; import static org.bytedeco.javacpp.opencv_highgui.CV_LOAD_IMAGE_GRAYSCALE; import static org.bytedeco.javacpp.opencv_highgui.imread; import static org.bytedeco.javacpp.opencv_imgproc.*; import org.slf4j.*; import pt.scanner.server.addon.*; import pt.scanner.server.data.*; import pt.scanner.server.main.Scanner; import pt.scanner.server.util.*; import static pt.scanner.server.util.Utils.*; /** * * @author brsantos */ public class Scanner implements Runnable { private static final Logger log = LoggerFactory.getLogger(Scanner.class); public static void main(String[] args) { Scanner scanner = new Scanner("../.."); while (true) { scanner.run(); } } private final String root; public Scanner(String root) { this.root = root; } @Override public void run() { try { final Mat grayImg = imread(String.format("%s/calibration/background.jpg", root), CV_LOAD_IMAGE_GRAYSCALE); Utils.showImage(grayImg, 1000); Mat binImg = new Mat(grayImg.size(), CV_8UC1, BLACK); threshold(grayImg, binImg, 100, 255, CV_THRESH_BINARY); Utils.showImage(binImg, 1000); ByteBuffer buffer = binImg.getByteBuffer(); IntStream.range(0, binImg.rows()).forEach(r -> { if (r < 2 || r > binImg.rows() - 3) { IntStream.range(0, binImg.cols()).forEach(c -> buffer.put(r * binImg.cols() + c, (byte) 255)); } else { buffer.put(r * binImg.cols(), (byte) 255); buffer.put(r * binImg.cols() + binImg.cols() - 1, (byte) 255); } }); // it is reusing the buffer - do not release Mat binImg2 = new Mat(binImg.rows(), binImg.cols(), CV_8UC1, new BytePointer(buffer)); Utils.showImage(binImg2, 1000); /* * Finding Contours */ MatVector contours = new MatVector(); findContours(binImg2, contours, CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE); log.info("{}", contours.size()); List<Contour> squares = LongStream.range(0, contours.size()) .filter(i -> Double.compare(contourArea(contours.get(i)), 20000.0) > 0) .mapToObj(i -> contours.get(i)) .sorted((c1, c2) -> Double.compare(contourArea(c1), contourArea(c2))).limit(30) .map(c -> new Contour(c, binImg2)).collect(Collectors.toList()); Mat bg = new Mat(binImg2.size(), CV_8UC1, BLACK); squares.forEach(c -> { log.info("contour area {}", contourArea(c.getContour())); MatVector v = new MatVector(1); v.put(0, c.getContour()); drawContours(bg, v, 0, new Scalar(255, 255, 255, 255)); Utils.showImage(bg, 1000); }); /* * Relating contours using nearness & position between lines */ squares.forEach(s -> s.joinLines()); Position.quadrants().forEach(pos -> { squares.forEach(c1 -> { Set<Line> l1 = c1.mainLines(pos); if (!l1.isEmpty()) { Line f1 = l1.iterator().next(); squares.forEach(c2 -> { Set<Line> l2 = c2.mainLines(pos); if (!l2.isEmpty()) { Line f2 = l2.iterator().next(); if (!c1.equals(c2) && !l1.isEmpty() && !l2.isEmpty() && ((f1.intersection(f2) != null && f1.angleDiff(f2) < 15) || (f1.intersection(f2) == null && f1.distancePerpendicular(f2.midPoint()) < 15))) { f1.addRelated(c2); f2.addRelated(c1); l1.add(f2); l2.add(f1); } } }); } }); }); Map<Position, Map<String, List<Line>>> meanLines = new HashMap<>(); Map<Position, Map<String, Line>> calibrationLines = new HashMap<>(); /* * Creating final main line for each position on each contour */ Position.quadrants().forEach(pos -> { meanLines.put(pos, new HashMap<>()); squares.stream().filter(c -> (c.mainLines(pos).size() > 0)).forEach(c -> { Line mean = Line.meanLine(c.mainLines(pos)).expandedLine(grayImg.size()); Set<Contour> related = c.mainLines(pos).stream().map(l -> l.getRelated()) .flatMap(con -> con.stream()).collect(Collectors.toSet()); mean.getRelated().addAll(related); String key = related.stream().map(r -> r.getIndex()).sorted().map(r -> r.toString()) .reduce((s1, s2) -> s1 + "|" + s2).get(); if (!meanLines.get(pos).containsKey(key)) { meanLines.get(pos).put(key, new ArrayList<>()); } log.debug("{} {} {}", pos, c, c.mainLines(pos)); c.mainLines(pos).clear(); c.mainLines(pos).add(mean); meanLines.get(pos).get(key).add(mean); }); Map<String, Line> linesPerContourSet = meanLines.get(pos).entrySet().stream() .collect(Collectors.toMap(e -> e.getKey(), e -> Line.meanLine(e.getValue()))); calibrationLines.put(pos, linesPerContourSet); linesPerContourSet.values().forEach(l -> line(grayImg, point(l.p0), point(l.p1), pos.color())); }); /* * Order contours using related main lines 1 - Find edge contours */ Contour topLeftContour = squares.stream() .sorted((c1, c2) -> (int) Math.round(c1.getCentroid().distance(new Coordinate(0, 0)) - c2.getCentroid().distance(new Coordinate(0, 0)))) .findFirst().get(); Contour topRightContour = squares.stream() .sorted((c1, c2) -> (int) Math.round(c1.getCentroid().distance(new Coordinate(grayImg.cols() - 1, 0)) - c2.getCentroid().distance(new Coordinate(grayImg.cols() - 1, 0)))) .findFirst().get(); Contour bottomLeftContour = squares.stream() .sorted((c1, c2) -> (int) Math.round(c1.getCentroid().distance(new Coordinate(0, grayImg.cols() - 1)) - c2.getCentroid().distance(new Coordinate(0, grayImg.cols() - 1)))) .findFirst().get(); Contour bottomRightContour = squares.stream() .sorted((c1, c2) -> (int) Math.round(c1.getCentroid() .distance(new Coordinate(grayImg.cols() - 1, grayImg.rows() - 1)) - c2.getCentroid().distance(new Coordinate(grayImg.cols() - 1, grayImg.rows() - 1)))) .findFirst().get(); /* * Order by region using related contours from main lines */ topLeftContour.sortRelated(0, 3, Position.RIGHT); topLeftContour.mainLines(Position.RIGHT).iterator().next().getRelated() .forEach(c -> c.sortRelated(c.getIndex(), 1, Position.BOTTOM)); topRightContour.sortRelated(12, 3, Position.LEFT); topRightContour.mainLines(Position.RIGHT).iterator().next().getRelated() .forEach(c -> c.sortRelated(c.getIndex(), 1, Position.BOTTOM)); bottomLeftContour.sortRelated(24, 1, Position.TOP); bottomLeftContour.mainLines(Position.TOP).iterator().next().getRelated().stream() .filter(c -> c.getIndex() == 25).findFirst().get().sortRelated(25, 1, Position.RIGHT); bottomRightContour.sortRelated(27, 1, Position.TOP); bottomRightContour.mainLines(Position.TOP).iterator().next().getRelated().stream() .filter(c -> c.getIndex() == 28).findFirst().get().sortRelated(28, 1, Position.LEFT); squares.forEach(c -> writeText(grayImg, 2.0, (int) FastMath.round(c.getCentroid().x), (int) FastMath.round(c.getCentroid().y), new Scalar(255, 0, 255, 0), c.getIndex().toString())); /* * Debug printing */ log.debug(calibrationLines.toString().replaceAll("\\], ", "]\n")); squares.forEach(s -> log.debug("{}", Corner.corners().stream().filter(cn -> s.getCorner(cn) != null) .collect(Collectors.toMap(cn -> cn, cn -> s.getCorner(cn))))); /* * calibration with laser on */ IntStream.range(-1, 128).mapToObj(i -> Integer.toString(i)) .map(i -> i.equals("-1") ? imread(String.format("%s/calibration/laser.jpg", root), CV_LOAD_IMAGE_GRAYSCALE) : imread(String.format("%s/imagens/capt%s.jpg", root, Strings.padStart(i, 3, '0')), CV_LOAD_IMAGE_GRAYSCALE)) .forEach(img -> { Utils.showImage(img, 1000); log.info("dilating"); Mat dilatedLaser = new Mat(img.size(), CV_8UC1, BLACK); Mat kernel = getStructuringElement(MORPH_RECT, new Size(3, 3), new opencv_core.Point(1, 1)); dilate(img, dilatedLaser, kernel); Utils.showImage(dilatedLaser, 1000); log.info("blurring"); Mat blurredLaser = new Mat(img.size(), CV_8UC3, BLACK); Mat blurredLaser2 = new Mat(img.size(), CV_8UC3, BLACK); medianBlur(img, blurredLaser, 15); Utils.showImage(blurredLaser, 1000); blur(blurredLaser, blurredLaser2, new Size(15, 15)); Utils.showImage(blurredLaser2, 1000); log.info("binary Image"); Mat binLaser = new Mat(img.size(), CV_8UC1, BLACK); threshold(blurredLaser, binLaser, 50, 255, CV_THRESH_BINARY); Utils.showImage(binLaser, 1000); log.info("Thinning"); Mat thinned = ZhangSuenThinning.thinning(binLaser); Utils.showImage(thinned, 1000); }); Thread.sleep(20000); } catch (RuntimeException | InterruptedException e) { log.error(e.getMessage(), e); } } }