Dear all,
I'm glad to introduce VXL-1.17.0 release, the source code and documentation of which
can be downloaded from I would like
to thank all guys for their hard work to keep VXL release pace back on track, and thank
Brendon for his patient guidance to me to make this release.
The primary changes made in this release are:
      - Moved the vxl photogrammetry library, vpgl, to core. vpgl supports a
        hierarchy of camera types including: projective, perspective and affine
        projections from 3-d to 2-d. Also included is a class for the
        epipolar relationship between cameras, vpgl_fundamental_matrix. An
        algorithm library is also provided, which implements standard
        operations such as bundle adjustment and backprojection.
      - Added transpose and conjugate_transpose methods to vnl_sparse_matrix
      - Added vnl_decnum: decimally represented arbitrary precision numbers;
        See vnl/examples/calculate.cxx for a useful application with vnl_decnum.
      - Removed broken vil_blob_finder, and replaced it with
        vil_blob, a collection of blob manipulation functions.
        The new API is more batch oriented rather than the old
      - Generalized vnl_sparse_lst_sqr_function and vnl_sparse_lm
        to allow for global parameters in the optimization.  Used
        this in vpgl_bundle_adjust to allow for shared intrinsic
        camera parameters in bundle adjustment.
      - Added a vbl_batch_multimap and vbl_compact_batch_multimap. These have
        vcl_multimap style interfaces, but with much lower memory allocation
        overheads when individual inserts and erases are not required.
      - Missing data support in vil. In particular image sampling can now return
        return NA when off edge of image. Compatible with vnl_na.
Lianqing Yu