A UAV SAR target detection method based on improved fukunage Koontz transform

With the development of synthetic aperture radar (SAR) technology, target detection algorithms in SAR images are confronted with difficulties, such as large scenes, complex environments, high resolution and poor real-time. The existing SAR target detection algorithms usually cannot meet the speed and accuracy of detection at the same time. Quadratic correlation filter (QCF), a simple but real-time object detection algorithm, is introduced to deal with the problem of target detection in SAR images. Fukunaga Koontz transform (FKT) is a useful method to design filters and coefficient matrix in QCF. In this paper, improved FKT method is proposed to detect the targets we want. First, the image is divided into several blocks to select regions of interest and improve the speed of our algorithm. Then, the kernel FKT (KFKT) method is used to detect targets in SAR images, which will make our algorithm more accurate. In order to prove the effectiveness of our experiment, the proposed method is compared with the classical Constant False Alarm Rate (CFAR) algorithm in SAR images and the original KFKT method. The simulation results show that our algorithm is superior to the other two methods in accuracy and rapidity.