【摘要】：在武器試驗(yàn)中，光電測(cè)量設(shè)備不僅要獲得目標(biāo)的外彈道軌跡、飛行軌跡等測(cè)試參數(shù)，同時(shí)還要獲得目標(biāo)的高精度飛行姿態(tài)參數(shù)、光學(xué)特性參數(shù)以及其它的常規(guī)測(cè)量參數(shù)。這些參數(shù)的取得依賴于所選擇的大面陣、高幀頻相機(jī)從不同角度獲取目標(biāo)的高速成像信息以及設(shè)備跟蹤的測(cè)量信息，這對(duì)高速圖像數(shù)據(jù)和同步測(cè)量信息的實(shí)時(shí)記錄提出了更高的要求，為此研究性能可靠、體積小、容量大，面向高速數(shù)據(jù)的存儲(chǔ)系統(tǒng)成為靶場(chǎng)測(cè)量急需解決的問(wèn)題，本文主要研究對(duì)象為面向大型光電經(jīng)緯儀的存儲(chǔ)系統(tǒng)，同樣適用于航測(cè)圖像、氣象雷達(dá)、遙感、爆炸實(shí)驗(yàn)、流體實(shí)驗(yàn)、碰撞檢測(cè)的存儲(chǔ)應(yīng)用中。 本文主要針對(duì)高速數(shù)據(jù)和低速存儲(chǔ)設(shè)備矛盾、圖像和異源信息同幀控制、嵌入式Linux操作系統(tǒng)移植、系統(tǒng)小型化、圖像數(shù)據(jù)的編碼糾錯(cuò)，存儲(chǔ)介質(zhì)的損耗均衡等問(wèn)題進(jìn)行研究，實(shí)現(xiàn)了系統(tǒng)高速數(shù)據(jù)的高可靠存儲(chǔ)和設(shè)備的小型化。 本文采用DM8168和FPGA相結(jié)合的嵌入式系統(tǒng)硬件平臺(tái)直接存儲(chǔ)方案，避免了非實(shí)時(shí)操作系統(tǒng)與計(jì)算機(jī)結(jié)構(gòu)的瓶頸限制，僅由兩個(gè)芯片完成了數(shù)據(jù)流的糾錯(cuò)編碼、采集、存儲(chǔ)和網(wǎng)絡(luò)傳輸控制，減少芯片級(jí)聯(lián)，降低硬件平均故障率，實(shí)現(xiàn)了系統(tǒng)的小型化。采用異源信息與圖像數(shù)據(jù)同幀追加策略，保證了追加異源信息與圖像數(shù)據(jù)的嚴(yán)格同幀存儲(chǔ)。 結(jié)合圖像的二維特點(diǎn)，本文設(shè)計(jì)了含有追加信息的實(shí)時(shí)圖像二維RS編碼糾錯(cuò)算法，在兩個(gè)維度上進(jìn)行RS編碼糾錯(cuò)，該算法提高了糾正數(shù)據(jù)突發(fā)錯(cuò)誤和隨機(jī)錯(cuò)誤的能力。同時(shí)還提出了采集容錯(cuò)和圖像二維特性容錯(cuò)方法，保證了存儲(chǔ)不丟幀，不誤幀，實(shí)現(xiàn)了數(shù)據(jù)糾錯(cuò)、幀糾錯(cuò)和行糾錯(cuò)，提高了數(shù)據(jù)存儲(chǔ)的可靠性。 本文對(duì)嵌入式Linux系統(tǒng)進(jìn)行移植，增加必要驅(qū)動(dòng)，裁剪掉系統(tǒng)無(wú)關(guān)模塊，保證了系統(tǒng)的穩(wěn)定運(yùn)行管理，系統(tǒng)對(duì)固態(tài)硬盤(pán)采用基于地址順序?qū)懭氲木夤芾�，順次遍歷固態(tài)硬盤(pán)所有的映射地址寫(xiě)入圖像數(shù)據(jù)，延長(zhǎng)了固態(tài)硬盤(pán)的使用壽命；對(duì)機(jī)械磁盤(pán)提出了采用直接控制物理地址的循道存儲(chǔ)的方案，減少尋道時(shí)間，提高存儲(chǔ)速度。 以DM8168為核心的嵌入式系統(tǒng)，進(jìn)行任務(wù)管理時(shí)，突破了硬盤(pán)文件系統(tǒng)的限制，將硬盤(pán)文件分配表等參數(shù)信息獨(dú)立記錄在存儲(chǔ)介質(zhì)外部FRAM中，該設(shè)計(jì)減少了Linux內(nèi)核對(duì)硬盤(pán)文件系統(tǒng)的操作開(kāi)銷(xiāo)，一方面便于機(jī)械磁盤(pán)數(shù)據(jù)恢復(fù)，另一方面解決了固態(tài)硬盤(pán)文件表需要損耗均衡的問(wèn)題，提高系統(tǒng)可靠性。 本文研究的存儲(chǔ)系統(tǒng)單元可實(shí)時(shí)穩(wěn)定存儲(chǔ)400MBps相機(jī)數(shù)據(jù)輸入，大大提高了存儲(chǔ)效率，，實(shí)現(xiàn)了設(shè)備的小型化設(shè)計(jì)，分塊二維RS編碼糾錯(cuò)的應(yīng)用使編碼糾錯(cuò)能力提高1.97倍，誤比特率為1.01E-11，嵌入式存儲(chǔ)系統(tǒng)實(shí)現(xiàn)了任務(wù)管理、數(shù)據(jù)遠(yuǎn)程網(wǎng)絡(luò)下載和系統(tǒng)的在線維護(hù)，根據(jù)工程實(shí)際應(yīng)用需要，采用針對(duì)高幀頻相機(jī)的數(shù)據(jù)源流分配技術(shù)，對(duì)存儲(chǔ)系統(tǒng)單元并行級(jí)聯(lián)后，可以滿足多領(lǐng)域的存儲(chǔ)應(yīng)用需要。
[Abstract]:In the weapon test, the photoelectric measuring equipment should not only obtain the testing parameters of the target's outer trajectory and flight trajectory, but also obtain the high-precision flight attitude parameters, optical characteristic parameters and other conventional measurement parameters of the target. The acquisition of these parameters depends on the large array selected, the high-speed imaging information of the target from different angles obtained by the high-frame-rate camera, and the measurement information tracked by the device. This puts forward higher requirements for real-time recording of high-speed image data and synchronous measurement information. Therefore, the research on high-speed data-oriented storage system with reliable performance, small size and large capacity has become an urgent problem to be solved in shooting range measurement. The main research object of this paper is the storage system for large-scale photoelectric theodolite, which is also suitable for the storage applications of aerial image, meteorological radar, remote sensing, explosion experiment, fluid experiment and collision detection. This paper mainly focuses on the contradictions between high-speed data and low-speed storage devices, image and heterogeneous information in the same frame control, embedded Linux operating system transplantation, system miniaturization, image data coding and error correction, loss equalization of storage media, and so on. The high reliability storage of high speed data and miniaturization of equipment are realized. In this paper, the direct storage scheme of embedded system hardware platform combined with DM8168 and FPGA is adopted to avoid the bottleneck limitation of non-real-time operating system and computer structure, and only two chips are used to complete the error-correcting coding and collecting of data stream. Storage and network transmission control can reduce chip concatenation, reduce the average failure rate of hardware, and realize the miniaturization of the system. The strategy of adding different source information and image data in the same frame is adopted to ensure the strict same frame storage of the additional source information and image data. Combined with the two-dimensional characteristics of the image, this paper designs a two-dimensional RS coding error correction algorithm for real-time images with additional information, and carries out RS coding error correction on two dimensions. The algorithm improves the ability of correcting data burst errors and random errors. At the same time, the methods of collecting fault tolerance and image two-dimensional characteristic fault tolerance are also proposed, which ensure that the storage can not lose frame and frame, realize data error correction, frame error correction and row error correction, and improve the reliability of data storage. In this paper, the embedded Linux system is transplanted, the necessary driver is added, and the system independent module is cut out, which ensures the stable operation and management of the system. The system adopts the balanced management based on address sequence writing to the solid state hard disk. After traversing all the mapping addresses of the solid state disk, the image data is written, which prolongs the service life of the solid state hard drive. This paper presents a scheme to control the physical address of the mechanical disk, which can reduce the searching time and improve the storage speed. In the embedded system with DM8168 as the core, the task management breaks through the limitation of the hard disk file system, and the parameter information, such as the hard disk file allocation table, is recorded independently in the external FRAM of the storage medium. The design reduces the operating overhead of the Linux kernel to the hard disk file system, on the one hand, it is convenient to recover the mechanical disk data, on the other hand, it solves the problem that the file table of the solid state disk needs to be balanced and improves the reliability of the system. The storage system unit studied in this paper can store the data input of 400MBps camera in real time and stably, which greatly improves the storage efficiency and realizes the miniaturization design of the equipment. The application of block two-dimensional RS coding and error correction improves the coding and error correction capability by 1.97 times. The bit error rate is 1.01Ex11. The embedded storage system realizes task management, data remote network download and on-line maintenance of the system. According to the requirements of engineering application, the data source and stream allocation technology for high frame-frequency camera is adopted. After parallel concatenation of storage system cells, it can meet the needs of multi-domain storage applications.
【學(xué)位授予單位】：中國(guó)科學(xué)院研究生院（長(zhǎng)春光學(xué)精密機(jī)械與物理研究所）
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2012
【分類(lèi)號(hào)】：TP333;TP391.41

【引證文獻(xiàn)】

相關(guān)期刊論文 前1條

1 李樂(lè)樂(lè);肖園;湯軻;;基于MSP430的簡(jiǎn)易阻抗測(cè)量?jī)x設(shè)計(jì)[J];科技視界;2013年02期

本文編號(hào)：2451168