發(fā)布時(shí)間：2018-09-18 11:57

【摘要】：隨著中國經(jīng)濟(jì)的快速發(fā)展,房地產(chǎn)市場(chǎng)熱度不減,各項(xiàng)基礎(chǔ)建設(shè)持續(xù)投入,對(duì)工程車輛的需求量也日益加大。工程車輛發(fā)揮舉足輕重作用的同時(shí),也伴隨著很多安全問題。工程車輛作業(yè)環(huán)境復(fù)雜,車身龐大,僅僅依靠后視鏡很難確保安全。而現(xiàn)代科學(xué)技術(shù)日新月異,許多工程車輛電子產(chǎn)品應(yīng)運(yùn)而生,將先進(jìn)的前沿技術(shù)應(yīng)用于傳統(tǒng)的工程車行業(yè)已是必然趨勢(shì)。全景環(huán)視影像系統(tǒng)能夠?qū)崟r(shí)提供車身周圍的俯瞰圖像,有效消除視野盲區(qū),為駕駛者提供了十分有效的輔助作用,研究該系統(tǒng)并將其應(yīng)用在工程車輛上具有很大的實(shí)用價(jià)值。本文分析了全景環(huán)視影像系統(tǒng)的背景和研究現(xiàn)狀,深入研究了該系統(tǒng)的關(guān)鍵技術(shù),根據(jù)工程車輛的實(shí)際特點(diǎn)給出了具體的實(shí)施方案。系統(tǒng)采用六個(gè)魚眼攝像頭對(duì)工程車輛四周進(jìn)行圖像采集,因采集的圖像帶有魚眼畸變,因此首先將圖像進(jìn)行畸變矯正,然后將矯正后的圖像經(jīng)透視變換轉(zhuǎn)換為俯視圖,最后將六幅圖像進(jìn)行拼接融合。在魚眼矯正算法中,本文分析了常用的幾種矯正算法,并通過仿真實(shí)驗(yàn)進(jìn)行效果對(duì)比,最終確定了使用標(biāo)定法進(jìn)行矯正。文中對(duì)標(biāo)定原理和方法做了詳細(xì)的介紹,并逐步實(shí)現(xiàn)了攝像頭的標(biāo)定,實(shí)驗(yàn)證明了標(biāo)定法矯正能夠滿足系統(tǒng)要求。圖像俯視變換階段,本文分析了其變換模型,確定了通過尋找角點(diǎn)來計(jì)算透視變換矩陣的方案,并給出了參考點(diǎn)選取的方法。在圖像匹配算法中,本文詳細(xì)介紹了匹配原理和SIFT算法的實(shí)現(xiàn)方法,并基于傳統(tǒng)的SIFT算法進(jìn)行了研究改進(jìn),提出了基于部分特征提取的改進(jìn)SIFT算法。該算法只在重合區(qū)域內(nèi)提取特征點(diǎn),大大減少了特征點(diǎn)數(shù)量和運(yùn)算時(shí)間,匹配成功率和效率都有所提升。文中給出了重合區(qū)域的確定方法,通過實(shí)驗(yàn)對(duì)比驗(yàn)證了改進(jìn)算法的有效性。在圖像融合算法中,本文研究了常用的幾種基于像素級(jí)的融合算法,通過仿真比對(duì),選取了漸入漸出的融合算法。漸入漸出融合算法在兩幅圖像亮度差別較大時(shí)無法完全消除拼接縫,圖像過渡不自然。本文針對(duì)此缺點(diǎn)對(duì)漸入漸出算法進(jìn)行了改進(jìn),對(duì)重合區(qū)域的平均灰度值進(jìn)行計(jì)算對(duì)比,將兩幅圖像灰度值調(diào)成相近后進(jìn)行融合。實(shí)驗(yàn)證明該改進(jìn)算法能有效消除拼接痕跡,使融合后圖像更加自然。本文研究了全景環(huán)視影像系統(tǒng)各環(huán)節(jié)算法,并改進(jìn)了魚眼矯正,圖像匹配和圖像融合算法,提升了全景圖的視覺效果,有效地減少了算法的復(fù)雜度,具有良好的應(yīng)用前景和價(jià)值。
[Abstract]:With the rapid development of Chinese economy, the heat of the real estate market is not decreasing, the investment of various infrastructure projects is continuing, and the demand for engineering vehicles is increasing day by day. Construction vehicles play a pivotal role, but also accompanied by a lot of safety problems. The working environment of engineering vehicles is complex and the body is huge. It is difficult to ensure safety only by rearview mirror. With the rapid development of modern science and technology, many electronic products of engineering vehicles emerge as the times require, and it is an inevitable trend to apply the advanced frontier technology to the traditional engineering vehicle industry. The panoramic circle view image system can provide the overlooking image around the car body in real time, effectively eliminate the blind area of the field of vision, and provide a very effective auxiliary function for the driver. It is of great practical value to study the system and apply it to the engineering vehicle. In this paper, the background and research status of panoramic circle view image system are analyzed, the key technology of the system is deeply studied, and the concrete implementation scheme is given according to the actual characteristics of engineering vehicles. The system uses six fish-eye cameras to collect images around the construction vehicle. Because the collected images have fish-eye distortion, the system first corrects the distortion of the images, and then converts the corrected images into the overlooking images by perspective transformation. Finally, the six images are stitched and fused. In the fish-eye correction algorithm, this paper analyzes several commonly used correction algorithms, and through simulation experiments to compare the results, finally determine the use of calibration method to correct. In this paper, the principle and method of calibration are introduced in detail, and the calibration of camera is realized step by step. The experimental results show that the calibration method can meet the requirements of the system. In the stage of image overhead transformation, this paper analyzes its transformation model, determines the scheme of calculating perspective transformation matrix by looking for corner points, and gives the method of selecting reference points. In the image matching algorithm, this paper introduces the matching principle and the realization method of SIFT algorithm in detail, studies and improves based on the traditional SIFT algorithm, and proposes an improved SIFT algorithm based on partial feature extraction. The algorithm only extracts feature points in the coincidence region, which greatly reduces the number of feature points and operation time, and improves the matching success rate and efficiency. In this paper, the method of determining the coincidence region is given, and the effectiveness of the improved algorithm is verified by experimental comparison. In the image fusion algorithm, this paper studies several commonly used fusion algorithms based on pixel level, and selects the gradually in and out fusion algorithm through simulation comparison. When the brightness difference between the two images is large, the image transition is not natural, and the splicing seam can not be completely eliminated. In order to solve this problem, this paper improves the incremental and gradual out algorithm, calculates and compares the average gray values of the coincidence region, and then adjusts the gray values of the two images to similar values and then fuses them together. Experiments show that the improved algorithm can effectively eliminate the stitching trace and make the fusion image more natural. In this paper, the algorithms of panoramic circle image system are studied, and the algorithms of fish-eye correction, image matching and image fusion are improved, the visual effect of panoramic image is improved, and the complexity of the algorithm is reduced effectively. It has good application prospect and value.
【學(xué)位授予單位】：中國礦業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TU603;TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 許超;聶詩良;;基于SURF和改進(jìn)漸入漸出法的圖像拼接算法[J];數(shù)字技術(shù)與應(yīng)用;2016年12期

2 何亞黎;袁義;;全景圖像拼接關(guān)鍵技術(shù)研究[J];信息化建設(shè);2016年03期

3 魏利勝;周圣文;張平改;孫駟洲;;基于雙經(jīng)度模型的魚眼圖像畸變矯正方法[J];儀器儀表學(xué)報(bào);2015年02期

4 韓迎輝;;基于改進(jìn)掃描線逼近的魚眼圖輪廓提取算法的研究[J];電子器件;2013年06期

5 王曉麗;戴華陽;余濤;謝東海;吳俁;;基于多分辨率融合的無人機(jī)圖像拼接勻色研究[J];測(cè)繪通報(bào);2013年06期

6 白廷柱;侯喜報(bào);;基于SIFT算子的圖像匹配算法研究[J];北京理工大學(xué)學(xué)報(bào);2013年06期

7 陳建明;曹永剛;;汽車電子安全技術(shù)的現(xiàn)狀及其發(fā)展策略[J];價(jià)值工程;2013年01期

8 蘇子孟;;工程機(jī)械行業(yè)面臨形勢(shì)和當(dāng)前主要工作[J];液壓氣動(dòng)與密封;2012年12期

9 韓昕;;工程機(jī)械行業(yè)發(fā)展趨勢(shì)概覽[J];今日工程機(jī)械;2012年21期

10 程勁波;;我國工程機(jī)械發(fā)展趨勢(shì)初探[J];科技信息;2011年16期

相關(guān)博士學(xué)位論文 前2條

1 唐晏;基于無人機(jī)采集圖像的植被識(shí)別方法研究[D];成都理工大學(xué);2014年

2 黃登山;像素級(jí)遙感影像融合方法研究[D];中南大學(xué);2011年

相關(guān)碩士學(xué)位論文 前10條

1 李星星;大比例尺多視角無人機(jī)遙感圖像拼接技術(shù)研究[D];杭州師范大學(xué);2015年

2 陳澤茂;基于全景視覺的汽車安全駕駛輔助系統(tǒng)的平臺(tái)設(shè)計(jì)與實(shí)現(xiàn)[D];華南理工大學(xué);2014年

3 楊楊;無人機(jī)航拍視頻圖像實(shí)時(shí)拼接軟件系統(tǒng)的設(shè)計(jì)與開發(fā)[D];北京工業(yè)大學(xué);2013年

4 劉新明;基于全景視覺的汽車輔助駕駛系統(tǒng)研究與實(shí)現(xiàn)[D];北京交通大學(xué);2013年

5 趙琦;基于魚眼鏡頭的全景圖像展開研究[D];長春理工大學(xué);2013年

6 周宇浩崴;基于DSP嵌入式平臺(tái)多路實(shí)時(shí)視頻拼接技術(shù)[D];上海交通大學(xué);2013年

7 韓文超;基于POS系統(tǒng)的無人機(jī)遙感圖像拼接技術(shù)研究與實(shí)現(xiàn)[D];南京大學(xué);2011年

8 張偉;魚眼圖像校正算法研究[D];南京郵電大學(xué);2011年

9 儀，

本文編號(hào)：2247831