CAJ | 학술논문

为了精确提取图像的细节特征，改善图像增强的效果，该文提出了一种基于形态学的可变权值自适应增强算法。通过构造扩展全方位多尺度结构元素，并进行top-hat变换，分解了图像不同方向不同尺度的细节。该文打破了传统形态学增强算法中各方向细节取均值融合的思想，实现了根据图像的局部灰度特性，动态调整各方向不同尺度细节的权值。在图像增强过程中，根据提取到的细节的结构化特征，构造相应的自适应增益函数，从而实现图像的自适应增强。实验结果表明，该算法能较好地利用图像的自相关性，比传统的形态学增强方法更有效地突出图像的细节信息，且能抑制噪声的放大。
위료정학제취도상적세절특정，개선도상증강적효과，해문제출료일충기우형태학적가변권치자괄응증강산법。통과구조확전전방위다척도결구원소，병진행top-hat변환，분해료도상불동방향불동척도적세절。해문타파료전통형태학증강산법중각방향세절취균치융합적사상，실현료근거도상적국부회도특성，동태조정각방향불동척도세절적권치。재도상증강과정중，근거제취도적세절적결구화특정，구조상응적자괄응증익함수，종이실현도상적자괄응증강。실험결과표명，해산법능교호지이용도상적자상관성，비전통적형태학증강방법경유효지돌출도상적세절신식，차능억제조성적방대。
In order to extract accurately the image details, and improve the effect of image enhancement, an adaptive image enhancement algorithm with variable weighted matching based on morphological is proposed. With this method, extension omni-directional multi-scale structure element is constructed, which is used to decompose image of different scale details in different direction through top-hat translation. The proposed algorithm brokes the idea of that the detail weighted in each direction is taken average in traditional morphology method, and adjusts the weight of the different detail direction based on the dynamic characteristic analysis of the local gray level. In the image enhancement process, according to the structured feature of extracted details, the corresponding adaptive gain function is constructed to realize the image adaptive enhancement. The experimental results show that, the algorithm can highlight more effective image details than the traditional morphological method of image enhancement by using the autocorrelation of image, and can suppress the noise in some extent.