复杂背景下彩色数字图像分割算法研究

Under complicated background color digital image segmentation algorithm research

摘要:

为了提高复杂背景下彩色数字图像分割的准确率,减少噪声对图像处理的干扰,尽可能的缩短运行时间,需要对复杂背景下彩色数字图像分割算法进行研究。当前图像分割算法在彩色数字图像分割的过程中,仅仅考虑了图像像素的亮度值,没有考虑其空间特征,存在计算的复杂性过大等缺陷,影响彩色数字图像处理效果。为此,提出了一种基于随机权重粒子群的复杂背景下彩色数字图像分割算法。该算法先采用多尺度均匀滤波方法,对复杂背景下彩色数据图像进行划分,其中包含噪声和不含噪声的像素点的亮度值、结构元素以及局部区域内的图像像素加权亮度密度特征。采用多段图分割获取彩色数字图像的优化分割,在平滑项中代入彩色数字图像梯度信息,对彩色数字图像分割结果中的弱边界进行剔除,从而实现准确的彩色数字图像分割。实验仿真证明,所提算法增加了彩色数字图像分割的对比度和信噪比,提高了复杂背景下彩色数字图像分割的准确性。

Abstract:

In the field of computer graphics and animation industry, the three dimensional animation image texture rendering technology has always been the main factor affecting the production cycle and cost of 3D animation image. So it is necessary to design the real-time rendering system of 3D animation image texture. The method of fast rendering based on special image texture. This method is different from the traditional ideas, not from the angle of calculation of polygon clipping to improve the rendering speed, but the integration of hardware processing technology of image texture, with fewer polygons model animation image texture rendering of high quality, in improving the image texture rendering speed and guarantee the quality of image texture rendering, but the method the calculation process is more complex. To this end, a method is proposed to design 3D animation image texture rendering system, the hardware design of the rendering system, the hardware of the system consists of image, image management, client node storage nodes and computing nodes. By drawing the corresponding coordinates of the rendered image model texture, the pixel can achieve the standard of the animation image texture rendering. The experimental results show that the proposed method can effectively reduce the cost of 3D animation image.