Research on Model Shape Optimization based on NURBS Modeling

Target recognition technology based on synthetic aperture radar (SAR) image plays an important role in several fields. Simulation images are needed as a supplement to achieve high recognition rate of SAR images. To expand the database based on the SAR image, it is necessary to optimize the shape of the target model to improve the similarity between the simulation image and the corresponding SAR image. This paper proposes a method of model shape optimization based on NURBS modeling. This method uses non-uniform rational B-spline (NURBS) to construct target model. Hummingbird optimization algorithm (HOA) is applied to optimize the shape of the target model, and complex wavelet structure similarity (CW-SSIM) is used as an evaluation index to compare the similarity between the simulation image and the target image. An example is given to show that this method can expand the image database effectively.