基于FPGA的仿造数码迷彩生成系统研究OA北大核心CSTPCD
Research on imitation digital camouflage generation system based on FPGA
为了解决传统仿造数码迷彩生成算法开发平台灵活性不足的问题,提出了一种基于现场可编程门阵列(FPGA)的仿造数码迷彩生成算法.采用直方图统计提取主要颜色信息,将背景图像马赛克化获得颜色分布信息,依据色差进行主色替换得到仿造数码迷彩的方法,取得了不同主色数量、色差计算权重、颜色量化数量下的数据.结果表明,颜色量化为3600 种后,提取出4 种主色作为仿造数码迷彩主色,根据色差主色替换时色彩权重3∶6∶1,生成的仿造数码迷彩具有良好的伪装效果,且基于FPGA实现该算法时系统占用资源少,处理 800 pixels×480 pixels分辨率的单帧彩色图像所需时间仅为22.21 ms,满足动态场景下生成仿造数码迷彩的实时性需求.该研究为主动伪装技术提供了参考.
To solve the problem of insufficient flexibility of the traditional counterfeit digital camouflage generation algorithm development platform,a counterfeit digital camouflage generation algorithm based on field-programmable gate array(FPGA)was proposed.The histogram statistics was used to extract main color information,the background image was mosaicked to obtain color distribution information,and the main colors based on color difference was replaced to obtain a method of imitating digital camouflage.Data were obtained with different main color numbers,color difference calculation weights,and color determination numbers.The results show that after determining 3600 colors,4 main colors are extracted as the main colors of the fake digital camouflage.When the main colors are replaced according to the color difference,the color weight is 3∶6∶1,the generated fake digital camouflage has a good encryption effect,and is based on FPGA implements this algorithm,the system occupies few resources.The time required to process a single frame color image with a resolution of 800 pixels×480 pixels is approximately 22.21 ms,which meets the real-time requirements for generating fake digital camouflage in dynamic scenes.This research provides a reference for active camouflage technology.
杜向坤;牛春晖;王晨;刘鑫
北京信息科技大学 仪器科学与光电工程学院,北京 100192,中国
图像处理仿造数码迷彩直方图统计颜色量化现场可编程门阵列
image processingimitation digital camouflagehistogram statisticscolor quantizationfield-programmable gate array
《激光技术》 2024 (005)
711-718 / 8
北京市自然科学基金资助项目(4224094);企业横向基金资助项目(GXTC19630439)
评论