FLANN Module API Reference

Complete documentation for the FLANN classes, methods, and enums in OpenCV5Sharp. Equivalent to the Official OpenCV Flann Documentation.

---

📦 Classes and Structs

FlannIndex

Inherits from: DisposableOpenCVObject

FLANN (Fast Library for Approximate Nearest Neighbors) index for performing fast nearest-neighbor searches. Wraps the native cv::flann::Index class and manages its lifecycle through DisposableOpenCVObject. The underlying native handle is released via flann_Index_Delete on disposal.

Constructors

• new FlannIndex()
  • Summary: Creates a new, empty FLANN index. The index must be populated later using Build() or Load() before performing searches.
  • Throws: OpenCVException if the underlying native call fails.
• new FlannIndex(Mat features, FlannIndexParams @params, IntPtr distType)
  • Summary: Creates a new FLANN index and immediately builds it from the given feature descriptors, index parameters, and distance type.
  • Parameter features: Matrix of feature descriptors (one row per feature) used to build the index.
  • Parameter params: Index parameters that control the algorithm and its configuration (e.g., KD-tree, KMeans, etc.).
  • Parameter distType: Native pointer specifying the distance metric to use when comparing features.
  • Throws: ArgumentNullException if features or params is null. ObjectDisposedException if features or params has been disposed. OpenCVException if the underlying native call fails.

Methods

• void Build(Mat features, FlannIndexParams @params, IntPtr distType)
  • Summary: Builds (or rebuilds) the FLANN index from the given feature descriptors, replacing any previously built index.
  • Parameter features: Matrix of feature descriptors (one row per feature) used to build the index.
  • Parameter params: Index parameters that control the algorithm and its configuration.
  • Parameter distType: Native pointer specifying the distance metric to use when comparing features.
  • Throws: ArgumentNullException if features or params is null. ObjectDisposedException if the index, features, or params has been disposed. OpenCVException if the underlying native call fails.
• void KnnSearch(Mat query, Mat indices, Mat dists, int knn, FlannSearchParams? @params)
  • Summary: Performs a K-nearest-neighbors search on the index. For each row in the query matrix, finds the knn closest feature vectors and writes their indices and distances into the output matrices.
  • Parameter query: Matrix of query feature vectors (one row per query point).
  • Parameter indices: Output matrix that will be filled with the indices of the nearest neighbors found.
  • Parameter dists: Output matrix that will be filled with the distances to the nearest neighbors found.
  • Parameter knn: Number of nearest neighbors to find for each query point.
  • Parameter params: Optional search parameters controlling the search behavior (e.g., number of checks). Pass null for defaults.
  • Throws: ArgumentNullException if query, indices, or dists is null. ObjectDisposedException if the index or any parameter has been disposed. OpenCVException if the underlying native call fails.
• int RadiusSearch(Mat query, Mat indices, Mat dists, double radius, int maxResults, FlannSearchParams? @params)
  • Summary: Performs a radius-based search on the index. For each row in the query matrix, finds all feature vectors within the specified distance radius, up to a maximum number of results.
  • Parameter query: Matrix of query feature vectors (one row per query point).
  • Parameter indices: Output matrix that will be filled with the indices of the neighbors found within the radius.
  • Parameter dists: Output matrix that will be filled with the distances to the neighbors found within the radius.
  • Parameter radius: The search radius; only neighbors within this distance are returned.
  • Parameter maxResults: Maximum number of neighbors to return per query point.
  • Parameter params: Optional search parameters controlling the search behavior. Pass null for defaults.
  • Returns: The number of neighbors found within the specified radius (capped by maxResults).
  • Throws: ArgumentNullException if query, indices, or dists is null. ObjectDisposedException if the index or any parameter has been disposed. OpenCVException if the underlying native call fails.
• void Save(string filename)
  • Summary: Saves the built FLANN index to a file so it can be reloaded later without rebuilding.
  • Parameter filename: Path to the file where the index will be saved.
  • Throws: ArgumentNullException if filename is null. OpenCVException if the underlying native call fails.
• bool Load(Mat features, string filename)
  • Summary: Loads a previously saved FLANN index from a file. The original feature descriptors must be provided since only the index structure is stored, not the feature data itself.
  • Parameter features: Matrix of feature descriptors that were used when the index was originally built.
  • Parameter filename: Path to the file from which to load the index.
  • Returns: true if the index was loaded successfully; false otherwise.
  • Throws: ArgumentNullException if features is null or filename is null. ObjectDisposedException if the index or features has been disposed. OpenCVException if the underlying native call fails.
• void Release()
  • Summary: Releases the internally held native FLANN index resources. After calling this method, the index object should not be used for searches.
  • Throws: OpenCVException if the underlying native call fails.
• IntPtr GetDistance()
  • Summary: Returns the distance metric type used by the built FLANN index.
  • Returns: A native pointer representing the distance type identifier used by this index.
  • Throws: OpenCVException if the underlying native call fails.
• IntPtr GetAlgorithm()
  • Summary: Returns the algorithm type used by the built FLANN index.
  • Returns: A native pointer representing the algorithm type identifier (e.g., KD-tree, KMeans) used by this index.
  • Throws: OpenCVException if the underlying native call fails.

---

FlannIndexParams

Inherits from: DisposableOpenCVObject

Represents index parameter settings for a FLANN index. Wraps the native cv::flann::IndexParams class. Used to configure the indexing algorithm (e.g., KD-tree, KMeans, composite, autotuned) when building a FlannIndex. The underlying native handle is released via flann_IndexParams_Delete on disposal.

Constructors

• new FlannIndexParams()
  • Summary: Creates a new FlannIndexParams instance with default parameter values.
  • Throws: OpenCVException if the underlying native call fails.

---

FlannSearchParams

Inherits from: DisposableOpenCVObject

Represents search parameter settings used when performing queries on a FLANN index. Wraps the native cv::flann::SearchParams class. Controls search behavior such as the number of recursive checks during a search. The underlying native handle is released via flann_SearchParams_Delete on disposal.

Constructors

• new FlannSearchParams()
  • Summary: Creates a new FlannSearchParams instance with default search parameter values.
  • Throws: OpenCVException if the underlying native call fails.

---

🔢 Enumerations

FlannFlannIndexType

Specifies the data type constants used by FLANN index parameters. These values correspond to OpenCV's internal type identifiers for FLANN parameter storage and mirror standard OpenCV depth types with additional FLANN-specific types.

Backing type: int

Member	Value	Description
FlannIndexType8u	0	8-bit unsigned integer type (corresponds to CV_8U).
FlannIndexType8s	1	8-bit signed integer type (corresponds to CV_8S).
FlannIndexType16u	2	16-bit unsigned integer type (corresponds to CV_16U).
FlannIndexType16s	3	16-bit signed integer type (corresponds to CV_16S).
FlannIndexType32s	4	32-bit signed integer type (corresponds to CV_32S).
FlannIndexType32f	5	32-bit floating-point type (corresponds to CV_32F).
FlannIndexType64f	6	64-bit floating-point type (corresponds to CV_64F).
FlannIndexTypeString	7	String parameter type for FLANN index configuration.
FlannIndexTypeBool	8	Boolean parameter type for FLANN index configuration.
FlannIndexTypeAlgorithm	9	Algorithm identifier parameter type for FLANN index configuration.
LastValueFlannIndexType	9	Sentinel value equal to the last defined type (FlannIndexTypeAlgorithm). Used for range validation.

---