摘要
雷达显示系统模拟中产生高质量雷达图像需要计算显示器分辨单元的全部像素,逐点计算效率低,严重影响模拟的实时性。利用像素着色器技术,将雷达图像分为3层独立绘制的纹理,并由硬件完成分层纹理的混合。降低了图像计算的复杂度,且充分利用了CPU与GPU(图像处理单元)的并行运算能力。应用结果表明,在雷达显示系统分辨力为1 280×1 024时,图像输出仍能稳定地保持在50 fps以上,能够很好地满足大分辨率雷达显示系统模拟的需求。
In the simulation of radar display system, radar image is generated by means of computing all the pixels＇s color, which is inefficient and has bad realtime performance. With the programmable render pipeline technique, it can evidently improve the efficiency of image generation, and delaminate the image into three layers that can be rendered independently, and then be mixed by hardware. The method reduces the complicity and takes full advantage of parallel computation in CPU and Graphic Processing Unit（GPU）. Its application shows that even as the radar displayer＇s resolution reaches 1280 × 1024, the image output rate still keeps more than 50 fps, which can meet the requirements of real-time simulation.
出处
《信息与电子工程》
2009年第3期180-183,共4页
information and electronic engineering
关键词
雷达模拟
雷达图像
像素着色器
余辉
radar simulation
radar image
pixel shader
persistence
