Moving point targets detection using cellular neural networks

The CNN-based Hough transform (HT) is presented. The HT is modified in order to be easily implemented by using the CNN. And the CNN-based HT is used to detect moving point targets under low signal-to-noise ratio (SNR) conditions. Owing to the inherent local, parallel, and analogy properties of the CNN and the strong robustness of the HT, our approaches are powerful and real time, which are illustrated by experimental simulation examples. A local line based filtering method is also presented, which is used to detect moving point targets under the conditions that noise or background is much stronger than signals