对应点生成 / Correspondence Generation

Functions
template<typename DataType , typename Signature , typename KNN >
std::vector< correspondence_t >	toolbox::pcl::generate_correspondences_knn (const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &src_cloud, const std::shared_ptr< std::vector< Signature > > &src_descriptors, const std::shared_ptr< std::vector< std::size_t > > &src_keypoints, const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &dst_cloud, const std::shared_ptr< std::vector< Signature > > &dst_descriptors, const std::shared_ptr< std::vector< std::size_t > > &dst_keypoints, float ratio=0.8f, bool mutual=true)
	对应关系生成策略 / Correspondence generation strategies
template<typename DataType , typename Signature >
std::vector< correspondence_t >	toolbox::pcl::generate_correspondences_brute_force (const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &src_cloud, const std::shared_ptr< std::vector< Signature > > &src_descriptors, const std::shared_ptr< std::vector< std::size_t > > &src_keypoints, const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &dst_cloud, const std::shared_ptr< std::vector< Signature > > &dst_descriptors, const std::shared_ptr< std::vector< std::size_t > > &dst_keypoints, float ratio=0.8f, bool mutual=true, bool parallel=false)
	快速生成对应关系的便捷函数（使用暴力搜索） / Convenience function for quick correspondence generation (using brute force)
size_t	toolbox::pcl::filter_correspondences_by_distance (std::vector< correspondence_t > &correspondences, float max_distance)
	过滤对应关系的辅助函数 / Helper function to filter correspondences
std::tuple< float, float, float, float >	toolbox::pcl::compute_correspondence_statistics (const std::vector< correspondence_t > &correspondences)
	计算对应关系的统计信息 / Compute statistics of correspondences
void	toolbox::pcl::print_correspondences (const std::vector< correspondence_t > &correspondences, size_t max_display=10)
	可视化对应关系的辅助信息 / Helper to visualize correspondence information

Detailed Description

Function Documentation

compute_correspondence_statistics()

std::tuple< float, float, float, float > toolbox::pcl::compute_correspondence_statistics ( const std::vector< correspondence_t > &  correspondences )

inline

计算对应关系的统计信息 / Compute statistics of correspondences

Parameters

correspondences

对应关系列表 / List of correspondences

Returns

包含均值、标准差、最小值、最大值的元组 / Tuple of mean, std, min, max

auto [mean, std, min, max] = compute_correspondence_statistics(correspondences);
std::cout << "距离统计 / Distance statistics: "
          << "均值 / mean=" << mean 
          << ", 标准差 / std=" << std 
          << ", 范围 / range=[" << min << ", " << max << "]" << std::endl;

filter_correspondences_by_distance()

size_t toolbox::pcl::filter_correspondences_by_distance ( std::vector< correspondence_t > &  correspondences, float  max_distance  )

inline

过滤对应关系的辅助函数 / Helper function to filter correspondences

Parameters

correspondences	[in/out] 对应关系列表 / List of correspondences
max_distance	最大允许距离 / Maximum allowed distance

Returns

被过滤掉的对应关系数量 / Number of filtered correspondences

// 过滤掉距离过大的对应关系 / Filter out correspondences with large distances
size_t filtered = filter_correspondences_by_distance(correspondences, 0.5f);
std::cout << "过滤掉 / Filtered out " << filtered << " 个对应关系 / correspondences" << std::endl;

generate_correspondences_brute_force()

template<typename DataType , typename Signature >

std::vector< correspondence_t > toolbox::pcl::generate_correspondences_brute_force	(	const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &	src_cloud,
		const std::shared_ptr< std::vector< Signature > > &	src_descriptors,
		const std::shared_ptr< std::vector< std::size_t > > &	src_keypoints,
		const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &	dst_cloud,
		const std::shared_ptr< std::vector< Signature > > &	dst_descriptors,
		const std::shared_ptr< std::vector< std::size_t > > &	dst_keypoints,
		float	ratio = 0.8f,
		bool	mutual = true,
		bool	parallel = false
	)

快速生成对应关系的便捷函数（使用暴力搜索） / Convenience function for quick correspondence generation (using brute force)

Template Parameters

DataType	数据类型 / Data type
Signature	描述子签名类型 / Descriptor signature type

Parameters

src_cloud	源点云 / Source point cloud
src_descriptors	源描述子 / Source descriptors
src_keypoints	源关键点索引 / Source keypoint indices
dst_cloud	目标点云 / Target point cloud
dst_descriptors	目标描述子 / Target descriptors
dst_keypoints	目标关键点索引 / Target keypoint indices
ratio	比率测试阈值（默认0.8） / Ratio test threshold (default 0.8)
mutual	是否启用双向验证（默认true） / Whether to enable mutual verification (default true)
parallel	是否启用并行计算（默认false） / Whether to enable parallel computation (default false)

Returns

对应关系列表 / List of correspondences

// 使用暴力搜索生成对应关系 / Generate correspondences using brute force
auto correspondences = generate_correspondences_brute_force<float, fpfh_signature_t<float>>(
    src_cloud, src_descriptors, src_keypoints,
    dst_cloud, dst_descriptors, dst_keypoints,
    0.8f, true, true);  // 启用并行 / Enable parallel

generate_correspondences_knn()

template<typename DataType , typename Signature , typename KNN >

std::vector< correspondence_t > toolbox::pcl::generate_correspondences_knn	(	const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &	src_cloud,
		const std::shared_ptr< std::vector< Signature > > &	src_descriptors,
		const std::shared_ptr< std::vector< std::size_t > > &	src_keypoints,
		const std::shared_ptr< toolbox::types::point_cloud_t< DataType > > &	dst_cloud,
		const std::shared_ptr< std::vector< Signature > > &	dst_descriptors,
		const std::shared_ptr< std::vector< std::size_t > > &	dst_keypoints,
		float	ratio = 0.8f,
		bool	mutual = true
	)

对应关系生成策略 / Correspondence generation strategies

不同的策略适用于不同的场景： Different strategies are suitable for different scenarios:

• 最近邻匹配：简单快速，但可能产生很多错误匹配 / Nearest neighbor: Simple and fast, but may produce many false matches
• 比率测试：Lowe's ratio test，有效减少错误匹配 / Ratio test: Lowe's ratio test, effectively reduces false matches
• 双向验证：要求匹配是相互的，进一步提高准确性 / Mutual verification: Requires mutual matching, further improves accuracy
• 几何一致性：检查匹配的几何关系，最可靠但计算量大 / Geometric consistency: Checks geometric relationships, most reliable but computationally intensive

// 配置对应点生成器 / Configure correspondence generator
correspondence_generator_t<float, fpfh_signature_t<float>, knn_t> gen;
 
// 宽松匹配（更多对应点，可能有错误） / Loose matching (more correspondences, possibly with errors)
gen.set_ratio(0.9f);
gen.set_mutual_verification(false);
 
// 严格匹配（更少但更可靠的对应点） / Strict matching (fewer but more reliable correspondences)
gen.set_ratio(0.7f);
gen.set_mutual_verification(true);
gen.set_distance_threshold(0.2f);

快速生成对应关系的便捷函数（使用KNN） / Convenience function for quick correspondence generation (using KNN)

Template Parameters

DataType	数据类型 / Data type
Signature	描述子签名类型 / Descriptor signature type
KNN	KNN搜索算法类型 / KNN search algorithm type

Parameters

src_cloud	源点云 / Source point cloud
src_descriptors	源描述子 / Source descriptors
src_keypoints	源关键点索引 / Source keypoint indices
dst_cloud	目标点云 / Target point cloud
dst_descriptors	目标描述子 / Target descriptors
dst_keypoints	目标关键点索引 / Target keypoint indices
ratio	比率测试阈值（默认0.8） / Ratio test threshold (default 0.8)
mutual	是否启用双向验证（默认true） / Whether to enable mutual verification (default true)

Returns

对应关系列表 / List of correspondences

// 使用KNN快速生成对应关系 / Quick correspondence generation using KNN
auto correspondences = generate_correspondences_knn<float, fpfh_signature_t<float>, kdtree_t<fpfh_signature_t<float>>>(
    src_cloud, src_descriptors, src_keypoints,
    dst_cloud, dst_descriptors, dst_keypoints,
    0.8f, true);

print_correspondences()

void toolbox::pcl::print_correspondences ( const std::vector< correspondence_t > &  correspondences, size_t  max_display = 10  )

inline

可视化对应关系的辅助信息 / Helper to visualize correspondence information

Parameters

correspondences	对应关系列表 / List of correspondences
max_display	最多显示的对应关系数量 / Maximum number of correspondences to display

// 打印前10个对应关系 / Print first 10 correspondences
print_correspondences(correspondences, 10);