Filter key points🔗
Discards points whose coordinates don’t fall inside a rectangular region of interest in world coordinates. The rectangular area is defined by a coordinate frame and a size, and may thus be warped by an affine transformation.
This tool is used in key point based matching as an optimization. It is usually preceded by cropping, which provides a quick way of extracting a sub-image. However, if the coordinate frame is not aligned to image axes, parts of the cropped image will be outside of the region of interest. The key point filtering tool is used as a postprocessing step to filter out key points extracted from such boundary areas.
The tool inspects all input points and removes those that fall outside of the region of interest. If dynamic input parameter matrices are provided, the corresponding rows in those will be removed as well.
points: A N-by-2 matrix that contains key point coordinates (x, y) as row vectors.
size: The size of the region of interest in world coordinates.
frame: A coordinate frame for the rectangular region of interest in the world coordinate system.
blockSize: The number of rows in each block of key points. This input provides a way for the key point detector to group key points belonging to separate objects.
dynamicInputCount: The number of dynamic input parameters.
matrixX: Dynamic input parameters (X ranges from 0 to dynamicInputCount - 1). They all accept an N-by-C matrix (C may be different in each input).
points: A M-by-2 matrix that contains the coordinates that fall inside the region of interest.
blockSize: The number of key points remaining in each block. A one-column matrix. If the blockSize input is not connected, this output will always be a 1-by-1 matrix that contains the number of rows in the points output.
matrixX: Dynamic output parameters(X ranges from 0 to dynamicInputCount - 1). Each will be a M-by-C matrix. The tool removes the corresponding rows from each output matrix.