基于ARM的移动场景多传感信号采集系统的研究与实现

发布时间：2018-06-03 00:41

本文选题：场景感知 + 日志记录　；参考：《电子科技大学》2017年硕士论文

【摘要】：生活场景感知是以某种技术手段对生活场景中的信号或事件进行捕获、识别的过程。生活场景感知通过记录生活中的细节信息,增强了用户的记忆能力。传统的事件记录方式为手写记录或电脑记录,这些手段虽然能够记录生活中的重要事件,但存在着记录时间不连续、操作时耗费大量人力的问题。现在由于可穿戴、移动多媒体技术的发展,通过电子手段能够实时地捕获生活场景中的细节信息。本课题研究的是基于多传感信号采集的生活场景感知系统,应用系统最终实现对生活场景中具体情境的感知与回放。具体研究工作包括:1.应用系统的硬件平台搭建。在分析了系统功能需求的基础上,搭建硬件平台,硬件平台包括ARM核心处理器模块、多路传感器,并在核心平台上对硬件驱动等进行了调试。2.设计系统的软件采集模块。采集模块的功能主要包括对位置信息的采集、光照信息的采集、音频信息的采集等。3.数据库的设计。设计数据库对信息进行存储,其中位置信息直接存储,光强信息简单分类后加以存储,声音信息特征提取后进行存储。4.声音识别与分类算法的设计。针对提取到的声音MFCC特征和基频特征,采用贝叶斯网络算法对声音信息进行分类。5.基于时间检索的实现。通过对时间维的检索,系统将同一时刻的多路信息相应的表征状态调入检索入口。6.系统感知模块的实现。实现位置识别、光强分类、声音分类模块、具体情境感知模块的设计。7.设计可视化的交互界面。由wxpython编写出可视化的交互界面,通过对检索入口的加载,界面中将显示对相关信息的检索结果。8.系统测试与实验。在系统设计的最后,对系统的性能进行了测试,接着设计了几组实验:声音识别的测试、位置对应的测试、光强识别的测试及信息检索的实验。生活场景感知应用系统实现了对生活场景具体情境感知的目标,即实现了信号的采集、存储、分类识别、感知、查询与显示等功能。
[Abstract]:Life scene perception is the process of capturing and recognizing the signals or events in the life scene by some technical means. Life scene perception enhances the user's memory ability by recording the details of life. The traditional method of recording events is handwritten records or computer records. Although these methods can record important events in life, they have the problem of discontinuous recording time and consuming a lot of manpower in operation. Now, with the development of wearable and mobile multimedia technology, the details of life scenes can be captured in real time by electronic means. This paper studies the life scene perception system based on multi-sensor signal acquisition. The application system finally realizes the perception and playback of the specific situation in the life scene. Specific research work includes: 1. The hardware platform of the application system is built. Based on the analysis of the functional requirements of the system, a hardware platform is built. The hardware platform includes ARM core processor module, multiple sensors, and the hardware driver is debugged on the core platform. The software acquisition module of the system is designed. The functions of the acquisition module mainly include the collection of position information, the collection of illumination information, the acquisition of audio information, etc. Database design. The design database stores the information, in which the position information is stored directly, the light intensity information is stored after simple classification, and the sound information feature is extracted and stored. Design of sound recognition and classification algorithm. According to the extracted MFCC features and fundamental frequency features, Bayesian network algorithm is used to classify the sound information. 5. Implementation of time-based retrieval. By retrieving the time dimension, the system transfers the representation state of the multi-channel information at the same time to the retrieval entry. 6. The realization of system perception module. Realization of position recognition, light intensity classification, sound classification module, specific situational awareness module design. 7. Design visual interactive interface. The visual interactive interface is written by wxpython. By loading the retrieval entry, the retrieval results of relevant information. 8. 8 will be displayed in the interface. System test and experiment. At the end of the system design, the performance of the system is tested, and then several experiments are designed: the test of sound recognition, the test of position correspondence, the test of light intensity recognition and the experiment of information retrieval. The life scene perception application system realizes the target of the life scene specific situation perception, namely realizes the signal collection, the storage, the classification identification, the perception, the inquiry and the display and so on.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP274.2

【参考文献】