基于FPGA的激光灭虫系统的研究

发布时间：2018-07-31 09:42

【摘要】：本文运用FPGA高速处理图像信息,根据目标检测技术和目标跟踪技术,达到对目标进行定位和跟踪的目的,这是应用视觉研究领域中的一个重要课题。随着当今时代计算机和信息科技的不断发展,机器视觉技术和目标识别算法都越来越普及的应用到了工业生产、医疗健康、军事项目以及农业生产中,机器视觉技术因为其高速度、高实时性、高准确率,而被广泛关注,其技术特点主要取决于目标识别算法的优良性,算法中重要的部分为视频图像采集技术和视频图像处理技术,这两个部分设计的复杂程度决定了整个系统功能实现效果,本文分析了很多资料的目标识别算法,研究设计了一种更加符合复杂背景环境的图像处理办法。本文采用了模块化的设计思想,有效降低了系统设计的相对复杂度,便于进行调试和修改；利用硬件语言设计体现了硬件电路中软件化的设计思路,提高了系统设计的灵活性与精确性；很好地发挥了FPGA可编程器件的集成度高、编程灵活的高效性,不仅有效地提升了系统的工作性能,而且将开发成本降至最低。本文首先用CMOS图像传感器采集一段视频数据,并用MATLAB将视频数据进行提取,然后对每个基本算法进行验证和修改,用ModelSim将图像处理算法进行实现,并进行功能仿真,最后用MATLAB将ModelSim仿真出来的数据再进行模拟对比,最后找出实验效果最好的一个算法,并对其进行改进。本文提出了一种应用于当今的图像处理技术上的目标识别算法,对其改进之后,并且实现在硬件FPGA上。该方案将整个系统分为六大模块,分别是图像采集模块、前期图像预处理模块、图像存储模块、图像读取模块、图像后处理模块、目标提取模块。通过对视频的严格处理,将其中运动的物体找到,并且将其分离出来,给出详细的位置信息,并控制激光进行击打或者跟踪。本次的实验结果表明,本文设计的各模块均能达到理想功能,基本接近功能仿真的效果,达到了预期的目标。
[Abstract]:In this paper, we use FPGA to process image information at high speed, and achieve the goal of target location and tracking according to target detection technology and target tracking technology, which is an important subject in the field of applied vision research. With the continuous development of computer and information technology, machine vision technology and target recognition algorithm are more and more widely used in industrial production, medical care, military projects and agricultural production. Machine vision technology has been widely concerned because of its high speed, high real-time and high accuracy. The technical characteristics of machine vision technology mainly depend on the excellence of target recognition algorithm. The important part of the algorithm is video image acquisition technology and video image processing technology. The complexity of these two parts determines the effect of the whole system function. In this paper, a lot of data target recognition algorithm is analyzed. In this paper, a new method of image processing is proposed, which is more suitable for complex background environment. This paper adopts the modular design idea, effectively reduces the relative complexity of the system design, facilitates debugging and modification, and embodies the software design idea in the hardware circuit by using the hardware language design. The flexibility and accuracy of the system design are improved, and the high integration degree of FPGA programmable devices and the high efficiency of programming flexibility are well brought into play, which not only effectively improves the working performance of the system, but also reduces the development cost to the lowest. This paper first uses CMOS image sensor to collect a piece of video data, then extracts the video data with MATLAB, then verifies and modifies each basic algorithm, implements the image processing algorithm with ModelSim, and simulates the function of the algorithm. Finally, MATLAB is used to simulate and compare the simulated data from ModelSim. Finally, an algorithm with the best experimental effect is found and improved. In this paper, a target recognition algorithm applied to image processing technology is proposed, which is improved and implemented on hardware FPGA. The whole system is divided into six modules, which are image acquisition module, pre-image preprocessing module, image storage module, image reading module, image post-processing module and target extraction module. Through the strict processing of the video, the moving objects are found, separated out, the detailed position information is given, and the laser is controlled to strike or track. The experimental results show that all the modules designed in this paper can achieve the ideal function, which are close to the effect of functional simulation and achieve the expected goal.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN791;TN249

【相似文献】