Add files via upload

CMakeLists.txt

@@ -0,0 +1,21 @@
+cmake_minimum_required(VERSION 3.5)
+project(extract_match_line)
+
+set(CMAKE_CXX_STANDARD 11)
+
+#OpenCV
+find_package(OpenCV)
+include_directories(OpenCV_INCLUDE_DIRS)
+
+#glogging
+find_package(Glog REQUIRED)
+
+#line_descriptor_opencvm
+include_directories(${PROJECT_SOURCE_DIR}/line_descriptor_opencvm/include)
+
+add_executable(extract_match_line main.cpp)
+target_link_libraries(extract_match_line
+        ${OpenCV_LIBS}
+        ${Glog_LIBRARIES}
+        ${PROJECT_SOURCE_DIR}/line_descriptor_opencvm/lib/liblinedesc.so
+        )

README.MD

@@ -0,0 +1,13 @@
+# 目标
+由于OpenCV3.4.6及以后版本将LSD线特征提取算法从代码中剔除，本库通过对OpenCV3.4.2及OpenCV-Contrib3.4.2的源代码中代码提取，提取出单独的LSD线特征提取算法并封装在cvm命名空间中。
+
+# 使用方式
+先进入line_descriptor_opencvm文件夹，然后
+```
+mkdir build
+cd build
+cmake ..
+make
+```
+之后会获得相应编译出的LSD算法的头文件和共享库文件
+然后对main函数进行cmake编译运行即可。就能得到两张图像的线特征提取，匹配的结果。

main.cpp

@@ -0,0 +1,69 @@
+#include <iostream>
+#include <vector>
+using std::cout;
+using std::endl;
+using std::vector;
+using std::string;
+
+#include <glog/logging.h>
+
+#include <opencv2/opencv.hpp>
+#include "descriptor_opencvm.hpp"
+
+#include <Eigen/Core>
+#include <Eigen/Geometry>
+
+int main(int argc, char** argv) {
+    //读入图像
+    google::InitGoogleLogging(argv[0]);
+    google::InstallFailureSignalHandler();
+    FLAGS_logtostderr=true;
+
+    double fx=481.2;
+    double fy=480;
+    double cx=319.5;
+    double cy=239.5;
+    //读入图像
+    cv::Mat color1 = cv::imread("../0.png",-1);     //先读入一帧RGB和depth
+    cv::Mat color2 = cv::imread("../50.png",-1);
+    if ( color1.data==nullptr || color2.empty()){
+        cout<<"wrong input rgb image"<<endl;
+    }
+    //初始化
+    cv::Ptr<cvm::line_descriptor::LSDDetector> lsd = cvm::line_descriptor::LSDDetector::createLSDDetector();
+    cv::Ptr<cvm::line_descriptor::BinaryDescriptor> lbd = cvm::line_descriptor::BinaryDescriptor::createBinaryDescriptor();
+
+    //提取图1的线特征
+    std::vector<cvm::line_descriptor::KeyLine> keylines1;
+    lsd->detect(color1, keylines1, 1.2, 1);  //提取线特征并计算描述子
+    LOG(INFO)<<"Extract line1 num = "<<keylines1.size();
+    cv::Mat ldesc1;
+    lbd->compute(color1, keylines1, ldesc1);     //计算特征线段的描述子
+    cv::Mat outlinesimg1;
+    cvm::line_descriptor::drawKeylines(color1,keylines1,outlinesimg1);
+    cv::imwrite("../colorline1.png",outlinesimg1);
+
+    //提取图2的线特征
+    std::vector<cvm::line_descriptor::KeyLine> keylines2;
+    lsd->detect(color2, keylines2, 1.2, 1);  //提取线特征并计算描述子
+    LOG(INFO)<<"Extract line2 num = "<<keylines2.size();
+    cv::Mat ldesc2;
+    lbd->compute(color2, keylines2, ldesc2);     //计算特征线段的描述子
+    cv::Mat outlinesimg2;
+    cvm::line_descriptor::drawKeylines(color2,keylines2,outlinesimg2);
+    cv::imwrite("../colorline2.png",outlinesimg2);
+    //匹配
+    cv::Ptr<cvm::line_descriptor::BinaryDescriptorMatcher> bdm=cvm::line_descriptor::BinaryDescriptorMatcher::createBinaryDescriptorMatcher();
+    std::vector<cv::DMatch> matches;
+    bdm->match(ldesc1,ldesc2,matches);
+    cv::Mat out_match_img;
+    LOG(INFO)<<"match num = "<<matches.size();
+
+    const std::vector<char> matchesMask(matches.size(),1);
+    cvm::line_descriptor::drawLineMatches(color1,keylines1,color2,keylines2,matches,out_match_img,
+            cv::Scalar::all(-1),cv::Scalar::all(-1),matchesMask,cvm::line_descriptor::DrawLinesMatchesFlags::DEFAULT);
+    cv::imwrite("../match_1_2.png",out_match_img);
+
+    std::cout << "Hello, World!" << std::endl;
+    return 0;
+}