To improve the automatic target recognition accuracy of SAR images and real-time performance

this study proposes a feature selection algorithm based on hybrid intelligent optimization for such images. First

a fractal feature is used to enhance an SAR image. An azimuth estimation method is then developed based on the image moment after image segmentation. Subsequently

the features of Zernike moment

Gabor wavelet coefficients

and gray level co-occurrence matrix are extracted from the original and the rectified images to form feature candidates. The genetic algorithm and the binary particle swarm optimization algorithm are combined to select features for SAR images. The effectiveness of the proposed algorithm is verified with the MSTAR database. Results demonstrate that the optimal feature sets can be generalized

thereby improving the target recognition rate and reducing recognition time.