從紅外圖像的特點出發，基于Canny算法進行了目標邊緣檢測。首先，對源圖像進行小波分解和重構，對圖像進行消噪，抑制噪聲對目標提取的影響。然后對消噪后的圖像用Canny算法進行邊緣檢測。這種邊緣檢測方法利用了小波消噪的優點，在抑制背景噪聲的同時保留了目標邊緣細節。實驗結果表明，這種結合方法優于單獨使用Canny算法。
關鍵詞：Canny 算法；紅外圖像；邊緣檢測；小波

Infrared target edge detection algorithm based on Canny algorithm Lian Jie, Han Chuanjiu, Pan Lu (Dept. of Communication and Information Engineering, Guilin Univ. of Electronic Technology, Guilin 541004) Abstract: According to the feature of infrared image, an improved Canny algorithm is proposed. The source image is decomposed and composed by wavelet, which could suppress the noise effect to targets. Then the denoised image is derived. And the edges of this image are detected using Canny operator. This edge detection method combines the advantages of both wavelet transform and Canny methods together, thus suppressing effectively the noises with the consecutive and clear edges kept up experimental results showed that the approach is superior to Canny method used alone.
Key words: Canny operator；Infrared image；Edge detection；Wavelet transform