基于离散分数布朗随机场模型的水下图像目标检测方法。该方法根据分形理论和水下图像的特点，以图像中每
个像素点为中心取窗口，计算在该窗口内的分形维数均值，将该均值作为中心像素的分形特征，然后根据分形维
数分布图确定分割阈值，从而实现对水下图像分割，并且通过将目标表面不同尺度下的灰度差分平均值进行归一
化处理，减少了用于表示不同尺度下的平均绝对值灰度差分的数据，从而提高算法检测效率(Underwater target detection method based on discrete fractional Brown random field model. According to the fractal theory and the characteristics of underwater images, the method takes every image in the image.

A pixel is taken as the center to take the window, and the mean value of the fractal dimension in the window is calculated. The mean value is taken as the fractal characteristic of the central pixel, and then according to the fractal dimension.

The number distribution map determines the segmentation threshold, so as to realize underwater image segmentation, and normalizing the average value of the gray scale difference under different scales of the target surface.

Chemical processing reduces the mean absolute value of the gray scale difference data at different scales, thus improving the detection efficiency of the algorithm.)