Java OpenCV FileStorage和Mat.push_back
我试图在Java中为KNN分类器实现this项目，即GenData.cpp（用C ++编写）。
我已经达到了这些代码并且卡住了：
matClassificationInts.push_back(intChar);
cv::FileStorage fsClassifications("classifications.xml", cv::FileStorage::WRITE);
fsClassifications << "classifications" << matClassificationInts;
fsClassifications.release();
在c ++中我们可以将整数传递给push_back（），但在Java中我得到错误：＆＃34; int无法转换为Mat＆＃34;。
所以，第一个问题是：如何将int传递给someMat.push_back（）？
第二个问题：如何在Java中实现FileStorage或将Mat编写为* .xml格式（并从* .xml中读取Mat）？
到目前为止，我的代码：
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Scanner;
import org.opencv.core.Core;
import static org.opencv.core.CvType.CV_32FC1;
import org.opencv.core.Mat;
import org.opencv.core.MatOfInt4;
import org.opencv.core.MatOfPoint;
import org.opencv.core.Rect;
import org.opencv.core.Scalar;
import org.opencv.core.Size;
import org.opencv.imgcodecs.Imgcodecs;
import org.opencv.imgproc.Imgproc;
import static org.opencv.imgproc.Imgproc.ADAPTIVE_THRESH_GAUSSIAN_C;
import static org.opencv.imgproc.Imgproc.CHAIN_APPROX_SIMPLE;
import static org.opencv.imgproc.Imgproc.RETR_EXTERNAL;
import static org.opencv.imgproc.Imgproc.THRESH_BINARY_INV;
public class genData {
private static final int
MIN_CONTOUR_AREA = 100,
RESIZED_IMAGE_WIDTH = 20,
RESIZED_IMAGE_HEIGHT = 30;
public static void main(String[] args) throws IOException {
System.loadLibrary(Core.NATIVE_LIBRARY_NAME);
Scanner keyboard = new Scanner(System.in);
boolean exit = false;
Mat imgTrainingNumbers;
Mat imgGrayscale = new Mat();
Mat imgBlurred = new Mat();
Mat imgThresh = new Mat();
Mat imgThreshCopy = new Mat();
ArrayList<MatOfPoint> ptContours = new ArrayList<MatOfPoint>();
MatOfInt4 v4iHierarchy;
Mat matClassificationInts = new Mat();
Mat matTrainingImagesAsFlattenedFloats = new Mat();
int[] intValidChars = { '0', '1', '2',
'A', 'B', 'C'}; //Here I did not make List<Integer>, because I can't pass char to Integer.
Arrays.sort(intValidChars); //for binary search
imgTrainingNumbers = Imgcodecs.imread("test.png"); //here Text on white paper.
if (imgTrainingNumbers.empty()) {
System.out.println("err");
return;
}
Imgproc.cvtColor(imgTrainingNumbers, imgGrayscale, Imgproc.COLOR_BGR2GRAY);
Imgproc.GaussianBlur(imgGrayscale, imgBlurred, new Size(5, 5), 0);
Imgproc.adaptiveThreshold(imgBlurred, imgThresh, 255, ADAPTIVE_THRESH_GAUSSIAN_C, THRESH_BINARY_INV, 11, 2);
/*
//imshow class implementation (found via google, works properly, but this block is commented for now)
Imshow im = new Imshow("imgThresh");
im.showImage(imgThresh);
imgThreshCopy = imgThresh.clone();
*/
Imgproc.findContours(imgThreshCopy, ptContours, new Mat(), RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
for (int i = 0; i < ptContours.size(); i++) {
if (Imgproc.contourArea(ptContours.get(i)) > MIN_CONTOUR_AREA) {
Rect boundingRect = Imgproc.boundingRect(ptContours.get(i));
Imgproc.rectangle(imgTrainingNumbers, boundingRect.tl(), boundingRect.br(), new Scalar(0, 0, 255), 2);
Mat matROI = imgThresh.submat(boundingRect.y, boundingRect.y + boundingRect.height, boundingRect.x, boundingRect.x + boundingRect.width);
Mat matROIResized = new Mat();
Imgproc.resize(matROI, matROIResized, new Size(RESIZED_IMAGE_WIDTH, RESIZED_IMAGE_HEIGHT));
/*
im.showImage(matROI);
im.showImage(matROIResized);
im.showImage(imgTrainingNumbers);
*/
String input = keyboard.nextLine();
int intChar = (int)input.charAt(0);
if (Arrays.binarySearch(intValidChars, intChar) >=0) {
/*
matClassificationInts.push_back(intChar);
//Here I'm getting an error.
*/
Mat matImageFloat = new Mat();
matROIResized.convertTo(matImageFloat, CV_32FC1);
Mat matImageFlattenedFloat = matImageFloat.reshape(1, 1);
matTrainingImagesAsFlattenedFloats.push_back(matImageFlattenedFloat);
}
}
}
//Here should go FileStorage stuff.
}
}
提前致谢。
附：使用OpenCV_310 + Java（不是JavaCV）
1 个答案:
答案 0 :(得分：0)
我自己做了解决方案。 非常脏，我猜，但它就是它。如果您知道如何对我的代码进行改进，我很高兴看到您的意见。
1）我的第一个问题是关于将一个int放入Mat（用于进一步制作* .xml）。我避免使用这种方法，并决定将int（实际上是Integer）放入List中。
Scanner keyboard = new Scanner(System.in);
String input = keyboard.nextLine();
int intChar = (int)input.charAt(0);
List<Integer> matClassificationInts = new ArrayList<Integer>();
if (Arrays.binarySearch(intValidChars, intChar) >=0) {
matClassificationInts.add(new Integer(intChar));
......
}
String dataImages = "";
for (Integer i : matClassificationInts) {
dataImages += i + " ";
}
我可以制作字符串（例如＆＃34; 49 48＆＃34;字符＆lt; - ＆gt;＆＃34; 1 0＆＃34; ints，例如）将其保存在* .xml中（参见下一段）。
2）第二个问题是关于从Mat提取数据并将其存储在* .xml中。好吧，通过C ++，我可以通过FileStorage来实现：
cv::FileStorage fsClassifications("classifications.xml", cv::FileStorage::WRITE);
fsClassifications << "classifications" << matClassificationInts;
fsClassifications.release();
但是Java OpenCV没有这样的功能，所以我遍历2d-array（Mat.rows（）和Mat.cols（））并通过get（）方法（Mat.get（row，col））来提取所需的数据 - 给出double的数组，array = 1的长度：
String dataClassifications = "";
for (int i = 0; i < matTrainingImagesAsFlattenedFloats.rows(); i++) {
for (int j = 0; j < matTrainingImagesAsFlattenedFloats.cols(); j++) {
double[] temp = matTrainingImagesAsFlattenedFloats.get(i, j);
dataClassifications += temp[0] + " ";
}
dataClassifications += "\n";
}
现在，关于将数据保存到* .xml：
我刚刚使用了 javafx.xml 和 org.wc3.dom 库。
制作了两个用于恢复DOM节点的函数：
private static Node getMatXML(Document doc, String option_id, String type_id, String rows, String cols, String dt, String data) {
Element elem = doc.createElement(option_id);
elem.setAttribute("type_id", type_id);
elem.appendChild(getMatXMLElement(doc,"rows", rows));
elem.appendChild(getMatXMLElement(doc, "cols", cols));
elem.appendChild(getMatXMLElement(doc, "dt", dt));
elem.appendChild(getMatXMLElement(doc, "data", data));
return elem;
}
private static Node getMatXMLElement(Document doc, String name, String value) {
Element node = doc.createElement(name);
node.appendChild(doc.createTextNode(value));
return node;
}
并使用这些函数创建* .xml：
Classifications.xml：
DocumentBuilderFactory icFactory_images = DocumentBuilderFactory.newInstance();
DocumentBuilder icBuilder_images;
try {
icBuilder_images = icFactory_images.newDocumentBuilder();
Document doc = icBuilder_images.newDocument();
Element mainRootElement = doc.createElement("opencv_storage");
doc.appendChild(mainRootElement);
mainRootElement.appendChild(getMatXML(doc, "classifications", "opencv-matrix", rowsImages, colsImages, "i", dataImages));
Transformer transformer = TransformerFactory.newInstance().newTransformer();
transformer.setOutputProperty(OutputKeys.INDENT, "yes");
DOMSource source = new DOMSource(doc);
String filename = "classifications.xml";
File file = new File(filename);
StreamResult console = new StreamResult(file); //(System.out)
transformer.transform(source, console);
} catch (Exception e) {
e.printStackTrace();
}
Images.xml：
DocumentBuilderFactory icFactory_classifications = DocumentBuilderFactory.newInstance();
DocumentBuilder icBuilder_classifications;
try {
icBuilder_classifications = icFactory_classifications.newDocumentBuilder();
Document doc = icBuilder_classifications.newDocument();
Element mainRootElement = doc.createElement("opencv_storage");
doc.appendChild(mainRootElement);
mainRootElement.appendChild(getMatXML(doc, "images", "opencv-matrix", rowsClassifications, colsClassifications, "f", dataClassifications));
Transformer transformer = TransformerFactory.newInstance().newTransformer();
transformer.setOutputProperty(OutputKeys.INDENT, "yes");
DOMSource source = new DOMSource(doc);
String filename = "images.xml";
File file = new File(filename);
StreamResult console = new StreamResult(file); //(System.out)
transformer.transform(source, console);
} catch (Exception e) {
e.printStackTrace();
}
因此，例如，生成的分类文件是：
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<opencv_storage>
<classifications type_id="opencv-matrix">
<rows>2</rows>
<cols>1</cols>
<dt>i</dt>
<data>49 48 </data>
</classifications>
</opencv_storage>
我测试了这张照片：
通过GenData.cpp（参见问题链接 - 第1行）和我的Java代码（完整代码见下文）。两个节目都给了我相同的结果：
对于Java OpenCV Imshow实现，您可以查看this link（不是我的）。
import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.FileReader;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
import java.util.Map;
import java.util.Scanner;
import org.opencv.core.Core;
import static org.opencv.core.CvType.CV_32FC1;
import org.opencv.core.Mat;
import org.opencv.core.MatOfInt4;
import org.opencv.core.MatOfPoint;
import org.opencv.core.Rect;
import org.opencv.core.Scalar;
import org.opencv.core.Size;
import org.opencv.imgcodecs.Imgcodecs;
import org.opencv.imgproc.Imgproc;
import static org.opencv.imgproc.Imgproc.ADAPTIVE_THRESH_GAUSSIAN_C;
import static org.opencv.imgproc.Imgproc.CHAIN_APPROX_SIMPLE;
import static org.opencv.imgproc.Imgproc.RETR_EXTERNAL;
import static org.opencv.imgproc.Imgproc.THRESH_BINARY_INV;
//XML - write.
import javax.xml.parsers.DocumentBuilder;
import javax.xml.parsers.DocumentBuilderFactory;
import javax.xml.transform.OutputKeys;
import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.stream.StreamResult;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.Node;
public class genData {
private static final int
MIN_CONTOUR_AREA = 100,
RESIZED_IMAGE_WIDTH = 20,
RESIZED_IMAGE_HEIGHT = 30;
public static void main(String[] args) throws IOException {
System.loadLibrary(Core.NATIVE_LIBRARY_NAME);
Scanner keyboard = new Scanner(System.in);
Mat imgTrainingNumbers;
Mat imgGrayscale = new Mat();
Mat imgBlurred = new Mat();
Mat imgThresh = new Mat();
Mat imgThreshCopy = new Mat();
ArrayList<MatOfPoint> ptContours = new ArrayList<MatOfPoint>();
MatOfInt4 v4iHierarchy = new MatOfInt4();
List<Integer> matClassificationInts = new ArrayList<Integer>();
Mat matTrainingImagesAsFlattenedFloats = new Mat();
int[] intValidChars = { '0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
'A', 'B', 'C', 'E', 'H',
'K', 'M', 'O', 'P', 'T',
'X', 'Y'};
Arrays.sort(intValidChars);
imgTrainingNumbers = Imgcodecs.imread("01.png");
if (imgTrainingNumbers.empty()) {
System.out.println("Error: file is not found");
return;
}
Imgproc.cvtColor(imgTrainingNumbers, imgGrayscale, Imgproc.COLOR_BGR2GRAY);
Imgproc.GaussianBlur(imgGrayscale, imgBlurred, new Size(5, 5), 0);
Imgproc.adaptiveThreshold(imgBlurred, imgThresh, 255, ADAPTIVE_THRESH_GAUSSIAN_C, THRESH_BINARY_INV, 11, 2);
Imshow im = new Imshow("imgThresh");
im.showImage(imgThresh);
imgThreshCopy = imgThresh.clone();
Imgproc.findContours(imgThreshCopy, ptContours, v4iHierarchy, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE);
for (int i = 0; i < ptContours.size(); i++) {
if (Imgproc.contourArea(ptContours.get(i)) > MIN_CONTOUR_AREA) {
Rect boundingRect = Imgproc.boundingRect(ptContours.get(i));
Imgproc.rectangle(imgTrainingNumbers, boundingRect.tl(), boundingRect.br(), new Scalar(0, 0, 255), 2);
Mat matROI = imgThresh.submat(boundingRect.y, boundingRect.y + boundingRect.height, boundingRect.x, boundingRect.x + boundingRect.width);
Mat matROIResized = new Mat();
Imgproc.resize(matROI, matROIResized, new Size(RESIZED_IMAGE_WIDTH, RESIZED_IMAGE_HEIGHT));
im.showImage(matROI);
im.showImage(matROIResized);
im.showImage(imgTrainingNumbers);
String input = keyboard.nextLine();
int intChar = (int)input.charAt(0);
if (Arrays.binarySearch(intValidChars, intChar) >=0) {
matClassificationInts.add(new Integer(intChar));
Mat matImageFloat = new Mat();
matROIResized.convertTo(matImageFloat, CV_32FC1);
Mat matImageFlattenedFloat = matImageFloat.reshape(1, 1);
matTrainingImagesAsFlattenedFloats.push_back(matImageFlattenedFloat);
}
}
}
String dataImages = "";
for (Integer i : matClassificationInts) {
dataImages += i + " ";
}
String dataClassifications = "";
for (int i = 0; i < matTrainingImagesAsFlattenedFloats.rows(); i++) {
for (int j = 0; j < matTrainingImagesAsFlattenedFloats.cols(); j++) {
double[] temp = matTrainingImagesAsFlattenedFloats.get(i, j);
dataClassifications += temp[0] + " ";
}
dataClassifications += "\n";
}
String rowsImages = String.valueOf(matClassificationInts.size());
String colsImages = "1";
String rowsClassifications = String.valueOf(matTrainingImagesAsFlattenedFloats.rows());
String colsClassifications = String.valueOf(matTrainingImagesAsFlattenedFloats.cols());
DocumentBuilderFactory icFactory_images = DocumentBuilderFactory.newInstance();
DocumentBuilder icBuilder_images;
try {
icBuilder_images = icFactory_images.newDocumentBuilder();
Document doc = icBuilder_images.newDocument();
Element mainRootElement = doc.createElement("opencv_storage");
doc.appendChild(mainRootElement);
mainRootElement.appendChild(getMatXML(doc, "classifications", "opencv-matrix", rowsImages, colsImages, "i", dataImages));
Transformer transformer = TransformerFactory.newInstance().newTransformer();
transformer.setOutputProperty(OutputKeys.INDENT, "yes");
DOMSource source = new DOMSource(doc);
String filename = "classifications.xml";
File file = new File(filename);
StreamResult console = new StreamResult(file); //(System.out)
transformer.transform(source, console);
} catch (Exception e) {
e.printStackTrace();
}
DocumentBuilderFactory icFactory_classifications = DocumentBuilderFactory.newInstance();
DocumentBuilder icBuilder_classifications;
try {
icBuilder_classifications = icFactory_classifications.newDocumentBuilder();
Document doc = icBuilder_classifications.newDocument();
Element mainRootElement = doc.createElement("opencv_storage");
doc.appendChild(mainRootElement);
mainRootElement.appendChild(getMatXML(doc, "images", "opencv-matrix", rowsClassifications, colsClassifications, "f", dataClassifications));
Transformer transformer = TransformerFactory.newInstance().newTransformer();
transformer.setOutputProperty(OutputKeys.INDENT, "yes");
DOMSource source = new DOMSource(doc);
String filename = "images.xml";
File file = new File(filename);
StreamResult console = new StreamResult(file); //(System.out)
transformer.transform(source, console);
} catch (Exception e) {
e.printStackTrace();
}
System.out.println("Finished.");
System.exit(0);
}
private static Node getMatXML(Document doc, String option_id, String type_id, String rows, String cols, String dt, String data) {
Element elem = doc.createElement(option_id);
elem.setAttribute("type_id", type_id);
elem.appendChild(getMatXMLElement(doc,"rows", rows));
elem.appendChild(getMatXMLElement(doc, "cols", cols));
elem.appendChild(getMatXMLElement(doc, "dt", dt));
elem.appendChild(getMatXMLElement(doc, "data", data));
return elem;
}
private static Node getMatXMLElement(Document doc, String name, String value) {
Element node = doc.createElement(name);
node.appendChild(doc.createTextNode(value));
return node;
}
}
相关问题
最新问题
- 我写了这段代码，但我无法理解我的错误
- 我无法从一个代码实例的列表中删除 None 值，但我可以在另一个实例中。为什么它适用于一个细分市场而不适用于另一个细分市场？
- 是否有可能使 loadstring 不可能等于打印？卢阿
- java中的random.expovariate()
- Appscript 通过会议在 Google 日历中发送电子邮件和创建活动
- 为什么我的 Onclick 箭头功能在 React 中不起作用？
- 在此代码中是否有使用“this”的替代方法？
- 在 SQL Server 和 PostgreSQL 上查询，我如何从第一个表获得第二个表的可视化
- 每千个数字得到
- 更新了城市边界 KML 文件的来源？