弱小目标低对比度图像增强算法研究

采用分段线性变换对白天亮背景下低对比度图像序列进行对比度增强,最关键的是分段点的选取。提出两种分段点确定方法:多尺度逼近利用目标和背景的灰度值在不同尺度空间中具有不同动态特性的特点,逐次逼近,通过找出目标和背景的过渡区域得到分段点;恒增强率方法确定一个固定的增强率,在直方图中由高到低找出满足该比率的灰度值作为分段点。两种方法均简单易行,试验证明它们能准确地对分段点进行定位,从而有效地增强了图像的对比度。

Enhancement of Low Contrast Image Contain Small Target

Selection of the subsection point is the key to enhance low contrast image under light background by subsection linear transformation. Present two methods to choose the subsection point:multiscale app roach (MSA) and constant enhance ratio (CER). The former is based on the princip le that target and background would express a different dynamic behavior in scale-space. It gets the subsection point in transition area between target and background by repetitious app roach. The latter defines a fixed enhance ratio, and find out the gray level that satisfied the ratio as the subsection point from high to low in the histogram. These two methods are simp le and fast. Experimental results prove that those methods can find the subsection point accurately, and enhance the contrast ratio of image effectively.