An efficient SAR target recognition algorithm based on contour and shape context

An efficient Recognition algorithm was proposed. The main idea of proposed algorithm is that the feature extraction module considers both the target and its shadow. The normalized Fourier series coefficients are used as the target contour feature, the spatial occurrence frequencies of both target and shadow pixels in several well constructed circular windows are used as shape context feature. The minimum distance classifier is used to give the final recognition result. Experiments on MSTAR data validate the effectiveness of proposed algorithm.

[1]  Gao Xin,et al.  SAR ATR based on Support Vector Machines and Independent Component Analysis , 2006, 2006 CIE International Conference on Radar.

[2]  L.M. Novak,et al.  An efficient multi-target SAR ATR algorithm , 1998, Conference Record of Thirty-Second Asilomar Conference on Signals, Systems and Computers (Cat. No.98CH36284).

[3]  Guan Jian,et al.  SAR ATR based on multi-subspaces of independent component analysis , 2007, 2007 1st Asian and Pacific Conference on Synthetic Aperture Radar.

[4]  Jitendra Malik,et al.  Shape matching and object recognition using shape contexts , 2010, 2010 3rd International Conference on Computer Science and Information Technology.

[5]  Saeid Belkasim,et al.  A novel curvature-based shape Fourier Descriptor , 2008, 2008 15th IEEE International Conference on Image Processing.

[6]  Ram M. Narayanan,et al.  Modeling of Target Shadows for SAR Image Classification , 2006, 35th IEEE Applied Imagery and Pattern Recognition Workshop (AIPR'06).

[7]  Michael Lee Bryant,et al.  Standard SAR ATR evaluation experiments using the MSTAR public release data set , 1998, Defense, Security, and Sensing.