视频图像去雪技术的FPGA实现(任务书,开题报告,论文17000字)
摘要
恶劣天气条件下对视频图像信息的检测对于气象中心公安交警部门等至关重要，特别是对空中，海上和地面交通管理的需求。在本文中，提出了一种基于FPGA的去除视频图像中雨雪的方法。室外监控系统的功能主要是在视频中对目标物体进行检测，对目标物体进行特征的提取以及行为的识别，基于这些功能可以设想到，如果在雪天的场景，必然会对这些操作造成很大影响，对于各个领域来讲，雪天的影响显得尤其明显。具体原因在于首先雪花的形状比较大，不透明度比较高，其次下落速度也比较慢，遮挡的面积也比较大，因此对在大雪天气的情况下，通过相关技术处理视频图像，获得图像质量清晰的视频成为当前视频图像处理领域的研究重点。为了在图像序列中将前景与背景分开，使用经典的高斯混合模型。前景模型用于检测雨雪，因为这些是动态天气现象。利用帧间差分或者背景差分的方法对视频图像中的雪花进行检测和清除。此外在检测和去除雪花之前要通过图像增强和图像平滑相关技术对图像进行实现处理，以便获得高质量的视频图像，为以后的工作做铺垫。该方法的应用很多，包括临界天气条件的检测，天气观测，视频监控系统的可靠性改进和雨水渲染。
关键词FPGA图像处理雨雪检测帧间差分中值滤波
Abstract
The detection of video image information under severe weather conditions is crucial for the meteorological department's public security traffic department, especially for air, sea and ground traffic management. In this paper, an FPGA-based method for removing rain and snow from video images is proposed. The function of the outdoor monitoring system is mainly to detect the target object in the video, extract the feature of the target object and identify the behavior. Based on these functions, it can be imagined that if it is in the snow scene, it will inevitably cause great impact on these operations. Impact, the impact of snow days is particularly evident for all relevant areas. The specific reason is that the shape of the snowflake is relatively large, the opacity is relatively high, and the falling speed is relatively slow, and the area of the occlusion is relatively large. Therefore, in the case of heavy snow weather, the video image is processed by the related technology to obtain a clear image quality. Video has become the focus of research in the field of video image processing. To separate the foreground from the background in the sequence of images, a classic Gaussian mixture model is used. The foreground model is used to detect rain and snow because these are dynamic weather phenomena. The snowflake in the video image is detected and cleared by means of interframe difference or background difference. In addition, the image is processed by image enhancement and image smoothing related techniques before the detection and removal of the snowflake, in order to obtain high-quality video images, paving the way for future work. The method is used in many applications, including detection of critical weather conditions, weather observation, reliability improvement of video surveillance systems and rainwater rendering.
Key wordsFPGA image processing, rain and snow detection, Interframe difference, median filterin
目录
第1章绪论 1
1.1 研究背景及意义 1
1.2 国内外研究现状 2
1.3 论文研究内容 3
1.3.1 研究思路 3
1.3.2 研究内容 4
第2章视频图像的处理 5
2.1 图像增强技术 5
2.1.1 灰度变换增强 5
2.1.2 直方图变换增强 6
2.2 图像去噪技术 10
2.2.1 邻域平均法 10
2.2.2 中值滤波法 12
第3章视频图像中雪花的检测和去除 15
3.1 背景的检测和提取 15
3.1.1 单高斯模型 15
3.1.2 混合高斯模型（MoG） 15
3.2 运动目标的检测 17
3.2.1 基于背景差分的方法 17
3.2.2 基于帧间差分的方法 18
第4章基于FPGA的系统体系设计 20
4.1 基于FPGA的帧间差分法的实现 20
4.2 基于FPGA的系统功能设计 23
4.3 系统流程介绍 26
第5章结果分析 27
第6章总结与展望 29
参考文献 30
致谢 33 |
