Function main

examples/cpp_example.cpp:108–167

std::string dataset2str(Dataset dataset_);

int main(int argc, char** argv)
{	
	// Parsing the flags
	gflags::ParseCommandLineFlags(&argc, &argv, true);

	printf("If you want to see the details of the fitting, e.g., time or estimation quality, turn on logging by starting the application as './SampleProject'\n");
	printf("Accepted flags:"); 
	printf("\n\t--problem-type {0,1,2} - The example problem which should run. Values: (0) Homography estimation, (1) Fundamental matrix estimation, (2) Essential matrix estimation. Default: 0");
	printf("\n\t--draw-results {0,1} - A flag determining if the results should be drawn and visualized. Default: 1");
	fflush(stdout);
	 
	/*
		This is an example showing how MAGSAC or MAGSAC++ is applied to homography or fundamental matrix estimation tasks.
		This implementation is not the one used in the experiments of the paper.
		If you use this method, please cite:
		(1) Barath, Daniel, Jana Noskova, and Jiri Matas. "MAGSAC: marginalizing sample consensus.", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2019.
		(2) Barath, Daniel, Jana Noskova, Maksym Ivashechkin, and Jiri Matas. "MAGSAC++, a fast, reliable and accurate robust estimator", Arxiv preprint:1912.05909. 2019.
	*/
	const double ransac_confidence = 0.99; // The required confidence in the results
	// The inlier threshold for visualization. This threshold is not used by the algorithm,
	// it is simply for selecting the inliers to be drawn after MAGSAC finished.
	const double drawing_threshold_essential_matrix = 3.00;
	const double drawing_threshold_fundamental_matrix = 1.00;
	const double drawing_threshold_homography = 1.00;

	switch (FLAGS_problem_type)
	{
		case 0:
			std::cout  << "Running homography estimation examples.";

			// Run homography estimation on the EVD dataset
			runTest(SceneType::HomographyScene, Dataset::extremeview, ransac_confidence, FLAGS_draw_results, drawing_threshold_homography);

			// Run homography estimation on the homogr dataset
			runTest(SceneType::HomographyScene, Dataset::homogr, ransac_confidence, FLAGS_draw_results, drawing_threshold_homography);
			break;
		case 1:
			std::cout  << "Running fundamental matrix estimation examples.";

			// Run fundamental matrix estimation on the kusvod2 dataset
			runTest(SceneType::FundamentalMatrixScene, Dataset::kusvod2, ransac_confidence, FLAGS_draw_results, drawing_threshold_fundamental_matrix);

			// Run fundamental matrix estimation on the AdelaideRMF dataset
			runTest(SceneType::FundamentalMatrixScene, Dataset::adelaidermf, ransac_confidence, FLAGS_draw_results, drawing_threshold_fundamental_matrix);

			// Run fundamental matrix estimation on the Multi-H dataset
			runTest(SceneType::FundamentalMatrixScene, Dataset::multih, ransac_confidence, FLAGS_draw_results, drawing_threshold_fundamental_matrix);
			break;
		case 2:
			std::cout  << "Running essential matrix estimation examples.";

			// Run essential matrix estimation on a scene from the strecha dataset
			runTest(SceneType::EssentialMatrixScene, Dataset::strecha, ransac_confidence, FLAGS_draw_results, drawing_threshold_essential_matrix);
			break;
		default:
			std::cerr << "Problem type " << FLAGS_problem_type << " is unknown. Valid values are 0,1,2.";
			break;
	}