Geometric Pattern Recognition System based on Statistical Parameters using Image Processing
Nishan Singh · 2015
The presented paper deals with the automatic visual inspection of the geometric patterns to recognise and classify. The scheme used for the classification of geometric patterns consists of four steps: 1) Pattern Generation 2) Making the pattern Rotation Invariant 3) Feature extraction and 3) Recognition and Classification. The pattern is generated by using any imaging tool e.g. by using paint-brush. The pattern is then made rotation independent by applying orthogonal transformation. After this, statistical features like, maximum radii in each quadrant, intercepts on each axis, perimeter, standard deviation, figure aspect and area, are computed from the image. The set of data of statistical parameters of each pattern so obtained is normalised to mean radius and stored for classification/categorization purposes.