平行多視點(diǎn)算法實(shí)時(shí)求取深度信息
REAL-TIME IMPLEMENTATION OF CALCULATING DEPTH INFORMATION BY IMAGE SEQUENCES ACQUIRED FROM THE PARALLEL MULTI-VIEWPOINTS
-
摘要: 引入了平行多視點(diǎn)的概念,從平行多視點(diǎn)獲取的多幅圖象中可以求得深度信息。通過(guò)利用圖象的FFT變換,將時(shí)域中很難解決的問(wèn)題變換到頻域中來(lái)解決,避免了復(fù)雜的匹配運(yùn)算,極大地簡(jiǎn)化了求取圖象點(diǎn)、面深度信息的問(wèn)題,并且提高了深度求取的實(shí)時(shí)性、全面性和準(zhǔn)確性。介紹了算法原理及特點(diǎn),并給出了實(shí)驗(yàn)結(jié)果及分析。
-
關(guān)鍵詞:
- 平行多視點(diǎn); 深度信息; 算法; 實(shí)時(shí)
Abstract: A new algorithm of calculating depth information by image sequences acquired from the parallel multi-viewpoints is discussed. By means of orthogonal transformation of image (in this paper, the authors use fast Fourier transformation), the matching problem, which is difficult to solve in time domain, is put into transformation field in frequency domain. -
Zhu X K, et al. AMSE Review, 1991, 19(3): 25-32.[2]Nasrabadi M. IEEE Trans, on PAMI, 1992, PAMI-14(5): 566-572.[3]Dhar M, Rangarajan K, Muiumdar J. Asia-Pacific Engineering Journal, Part A [Electrical Eagineering], 1992. 2(2):217-231.[4]Sparr G. Image and Vision Computing,[5]92, 10(10): 683-688. -
計(jì)量
- 文章訪問(wèn)數(shù): 2053
- HTML全文瀏覽量: 115
- PDF下載量: 447
- 被引次數(shù): 0
下載:
