Research on Boosting Canny Edge Detection Performance with DCT and NLM

Low altitude aerial target detection is essential for military, civil aviation, and unmanned aerial system applications. However, conventional Canny edge detection algorithms have limitations that impede performance in noisy environments. Specifically, they are prone to losing faint edge information during noise filtering owing to poor adaptability with fixed parameters. To address these limitations, this study presents a modified Canny edge detection algorithm for low altitude aerial target edge detection. The proposed method substitutes the traditional Gaussian filter in the Canny algorithm with a combined discrete cosine transform (DCT) and non-local means (NLM) filtering approach. Comparative experimental results demonstrate that this novel technique exhibits excellent capabilities in edge detection and noise suppression under low altitude aerial target conditions. The findings provide an innovative solution for robust detection of low altitude aerial targets in real-world noisy environments, offering promise in advancing military, aviation, and unmanned aerial system applications.