The MASM Forum Archive 2004 to 2012

Quote
void cvExtractSURF(const CvArr*  image, const CvArr*  mask, CvSeq**  keypoints, CvSeq**  descriptors, CvMemStorage*  storage, CvSURFParams params)¶

    Extracts Speeded Up Robust Features from an image.
    Parameters:   

        * image – The input 8-bit grayscale image
        * mask – The optional input 8-bit mask. The features are only found in the areas that contain more than 50 % of non-zero mask pixels
        * keypoints – The output parameter; double pointer to the sequence of keypoints. The sequence of CvSURFPoint structures is as follows:

    typedef struct CvSURFPoint
    {
       CvPoint2D32f pt; // position of the feature within the image
       int laplacian;   // -1, 0 or +1. sign of the laplacian at the point.
                        // can be used to speedup feature comparison
                        // (normally features with laplacians of different
                // signs can not match)
       int size;        // size of the feature
       float dir;       // orientation of the feature: 0..360 degrees
       float hessian;   // value of the hessian (can be used to
                // approximately estimate the feature strengths;
                        // see also params.hessianThreshold)
    }
    CvSURFPoint;

    Parameters:   

        * descriptors – The optional output parameter; double pointer to the sequence of descriptors. Depending on the params.extended value, each element of the sequence will be either a 64-element or a 128-element floating-point ( CV_32F ) vector. If the parameter is NULL, the descriptors are not computed
        * storage – Memory storage where keypoints and descriptors will be stored
        * params – Various algorithm parameters put to the structure CvSURFParams:

    typedef struct CvSURFParams
    {
       int extended; // 0 means basic descriptors (64 elements each),
                     // 1 means extended descriptors (128 elements each)
       double hessianThreshold; // only features with keypoint.hessian
             // larger than that are extracted.
                     // good default value is ~300-500 (can depend on the
             // average local contrast and sharpness of the image).
                     // user can further filter out some features based on
             // their hessian values and other characteristics.
       int nOctaves; // the number of octaves to be used for extraction.
                     // With each next octave the feature size is doubled
             // (3 by default)
       int nOctaveLayers; // The number of layers within each octave
             // (4 by default)
    }
    CvSURFParams;

    CvSURFParams cvSURFParams(double hessianThreshold, int extended=0);
             // returns default parameters

The function cvExtractSURF finds robust features in the image, as described in Bay06 . For each feature it returns its location, size, orientation and optionally the descriptor, basic or extended. The function can be used for object tracking and localization, image stitching etc.
http://opencv.willowgarage.com/documentation/c/feature_detection.html