水下爆炸气泡图像边界提取方法

Extraction Method of Underwater Explosion Bubble Image Boundary

为准确提取水下爆炸气泡图像边界,尝试采用在边界识别中常用的canny算子边缘检测方法,但结果并不理想。在此基础上提出了先将灰度图像转换成索引图像,采用二维离散小波分解方法小波变换方法对索引图像进行分解,针对分解的图像轮廓部分主要体现在低频部分的特点对低频部分进行增强处理,对高频部分进行衰减处理,然后对处理后的图像进行canny算子边缘检测。结果表明:小波增强后的canny算子边缘检测方法处理出来的气泡图像边界清晰,可作为进一步分析研究的依据。

In order to obtain the image boundary of underwater explosion bubble,the edge detection based on canny operator was commonly used in the boundary recognition,while the result was not good enough. Base on the canny operator method,the gray image was transformed into index image and then decomposed by the two-dimensional discrete wavelet decomposition. Commonly,the low-frequency part mainly represented the picture contour,and enhanced low-frequency part and decayed the high-frequency part. Therefore,the canny operator method was used to detect the processed image. The result showed that the image boundary of processed bubble image was detected clearly by the wavelet enhancement canny operator method.