本文選題：SLR + 圖像能量累積�。� 參考：《中國地震局地震研究所》2013年碩士論文

【摘要】：空間碎片又稱太空垃圾，因其對太空環(huán)境造成極大的危害，而受到各國重視，目前對空間碎片采取的主要措施就是觀測。SLR(Satellite Laser Ranging)技術(shù)是現(xiàn)代空間大地測量最先進技術(shù)之一。在SLR觀測中，帶激光反射棱鏡的觀測對象稱為合作目標，像空間碎片等不帶有反射鏡的觀測對象稱為非合作目標。本論文要研究空間碎片SLR圖像跟蹤技術(shù)，由于SLR觀測技術(shù)是被動跟蹤，故圖像跟蹤技術(shù)的核心是目標識別，研究的主要內(nèi)容是對SLR圖像增強器上采集下的非合作目標圖像進行目標識別。 本研究主要分為5個部分：圖像采集，圖像去噪，圖像增強，目標識別和目標分析。 第一部分：圖像采集是基于MATLAB中ImageAcquisition Toolbox工具實現(xiàn)的，由于MATLAB自帶圖像采集模塊，只需設(shè)置采集參數(shù)。 第二部分：圖像去噪部分是研究的重點，研究中針對衛(wèi)星運行多樣性，選取目標易識別和目標不易識別兩種情況下的觀測圖像作為研究對象，并分別采用形態(tài)學去噪中的開運算、閉運算，空間域去噪法中的中值濾波、均值濾波、自適應(yīng)性濾波，頻域去噪法中的離散余弦去噪對樣本進行單幀圖像的去噪實驗，效果均不明顯，后采用有限序列圖像累加的方式進行圖像去噪，效果明顯。 第三部分：圖像增強采用laplace算子、sobel算子、prewitt算子和log算子進行實驗，均達不到實驗?zāi)康�，后采用負相變換結(jié)合線性運算，實現(xiàn)目標周圍灰度差增大。 第四部分：目標識別采用閾值分割法及邊緣檢測法中的基于sobel算子和robinson算子的邊緣分割對圖像進行目標識別，由于噪聲干擾過多，對于小目標圖像識別有困難，經(jīng)過有效去噪后，進行簡單的閾值分割，實現(xiàn)目標識別，，后利用目標形態(tài)進行目標篩選，精確識別目標。 第五部分：目標分析是標識和計算目標位置，最后可返回目標坐標。 最終通過研究實現(xiàn)了對SLR觀測圖像的目標識別，為今后SLR跟蹤觀測技術(shù)的發(fā)展提供參考。
[Abstract]:Space debris is also known as space garbage. Because of its great harm to the space environment, it has been paid much attention by all countries. The main measure to take space debris is the observation of.SLR (Satellite Laser Ranging) technology is one of the most advanced technologies in modern space geodetic. In the SLR observation, the object of observation with a laser reflective prism is called cooperation. The object, such as space debris and non reflecting mirrors, is called non cooperative target. In this paper, we should study space debris SLR image tracking technology. Because SLR observation technology is passive tracking, the core of image tracking technology is target recognition. The main content of the research is the non cooperative target image acquisition on the SLR image intensifier. Line target recognition.
This study is mainly divided into 5 parts: image acquisition, image denoising, image enhancement, target recognition and target analysis.
The first part: image acquisition is based on ImageAcquisition Toolbox tool in MATLAB. Since MATLAB has its own image acquisition module, it only needs to set acquisition parameters.
The second part: the image denoising part is the focus of the study. In the study, the diversity of the satellite operation is selected, and the observation images under two conditions are selected and the target is easy to recognize and the target is not easily recognized as the research object, and the open operation in the morphological denoising, the closed operation, the mean filtering, the mean filtering and the self-adaptive in the space domain denoising method are adopted respectively. Filtering, the discrete cosine de-noising in the frequency domain denoising method is used to denoise the single frame of the sample, and the effect is not obvious. Then the image denoising is carried out with the finite sequence image accumulation.
The third part: the image enhancement uses the Laplace operator, the Sobel operator, the Prewitt operator and the log operator to carry out the experiment, which can not reach the experimental purpose, and then the negative phase transformation is combined with the linear operation, and the gray scale difference around the target is increased.
The fourth part: the target recognition uses the threshold segmentation method and the edge detection method based on the Sobel operator and the Robinson operator to recognize the image. Because of the excessive noise interference, it is difficult for the small target image recognition. After effective denoising, a simple threshold segmentation is carried out to realize the target recognition, and then the target shape is used. The state carries out the target screening and accurately identifies the target.
The fifth part: target analysis identifies and calculates the location of the target, and finally returns the target coordinates.
Finally, the target recognition of SLR observation images is realized, which provides reference for the development of SLR tracking technology in the future.
【學位授予單位】：中國地震局地震研究所
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：P225.2;X738

【參考文獻】

相關(guān)期刊論文 前10條

1 林有作,成麗波;基于小波分析的數(shù)字圖像噪聲消除[J];長春理工大學學報;2004年01期

2 李怡勇;沈懷榮;李智;;空間碎片環(huán)境危害及其對策[J];導(dǎo)彈與航天運載技術(shù);2008年06期

3 于成忠;朱駿;袁曉輝;;基于背景差法的運動目標檢測[J];東南大學學報(自然科學版);2005年S2期

4 王帥;趙輝;周臨川;;基于Mean-Shift的目標跟蹤算法[J];電氣電子教學學報;2011年01期

5 張澤旭,李金宗,李寧寧;基于光流場分割和Canny邊緣提取融合算法的運動目標檢測[J];電子學報;2003年09期

6 何偉,晉兆虎,張玲;一種改進的利用背景檢測弱小目標的方法[J];重慶大學學報(自然科學版);2005年07期

7 胡鵬;徐會艷;;基于Matlab的圖像去噪算法的研究與實現(xiàn)[J];福建電腦;2009年12期

8 王姣斐;王雙喜;;基于MATLAB軟件的圖像去噪方法比較[J];甘肅農(nóng)業(yè)大學學報;2011年04期

9 薛陳;朱明;陳愛華;;魯棒的基于改進Mean-shift的目標跟蹤[J];光學精密工程;2010年01期

10 屈有山,田維堅,李英才;基于并行隔幀差分光流場與灰度分析綜合算法的運動目標檢測[J];光子學報;2003年02期

本文編號：1915393