无人机机载视频去雾系统设计

Design of the Airborne Video Demisting System for UCAV

摘要:

针对无人机机载雾天获取图像降质的问题，设计基于嵌入式模块化视频去雾系统。系统选用TS1601芯片作为视频去雾处理的核心，设计数字视频BT.656/BT.1120隔行和逐行处理接口，iMX6嵌入式系统完成去雾参数化控制、H.264视频压缩处理和组帧处理、外围接口设计和控制指令响应等设计，完成去雾参数化处理、去雾控制功能、内循环等功能模块设计流程和实现。实验测试结果表明，该系统满足无人机机载视频去雾处理的需求，具有功耗小、易实现的特点和适应性强等特点，满足无人机机载视频实时性去雾处理要求。在客观评价中，本系统和相关文献方法对典型含雾图像处理后并分别采用TenenGrad函数、方差函数、平均梯度梯度函数等三种清晰度评价函数并做归一化处理进行评价，其中对于含雾图像3处理后的TenenGrad函数归一化分别提高47%、1.40%、12.80%，方差函数归一化分别提高12.50%、9.20%、11.20%，平均梯度梯度函数归一化分别提高53.20%、3.60%、8.90%，整体运算时间提高4.74倍、5.41倍和5.46倍。

Abstract:

Arming at the problem of image degradation during the UCAV capture the image for the interested target in the foggy day,we propose a system for video demisting based on embedded technology. The system adopts TS1601 as the main unit for the whole hardware system to complete demisting parameterization control，H.264 video compression processing and framing processing,peripheral interface design and control command response and other function. This paper introduces the design of the system’s design framework,software design frame work and software modules.The experimental results indicate that the system can fulfill the demand of video demisting in real time and have advantages such as low power consumption，easy implement. It can achieve the demand of the airborne demisting application.In the objective evaluation, our method and related references methods have used three kinds of definition evaluation functions (such as TenenGrad function, variance function and average gradient function) to evaluate performance for the typical misting image.For the image 3, the normalization of TenenGrad function after the demisting is increased by 47%, 1.40% and 12.80%, and the normalization of variance function is increased by 12.50% , 9.20%, and 11.20%, the normalization of the average gradient function increased by 53.20%, 3.60%, 8.90% respectively. And the overall operation time of our method is increased by 4.74 times, 5.41 times and 5.46 times.