本文關(guān)鍵詞： 敏感區(qū)域 三角網(wǎng)格 魚眼變換 圖像縮放　出處：《微電子學與計算機》2017年10期 　論文類型：期刊論文

【摘要】：為了解決傳統(tǒng)內(nèi)容感知圖像縮放技術(shù)縮放效果的不確定性,且方法不適用于多光譜遙感圖像的問題,提出了一種三角網(wǎng)格與魚眼變換相結(jié)合的圖像縮放方法.首先,根據(jù)敏感目標的特征識別標注出敏感目標區(qū)域,然后,對標注出的敏感區(qū)域內(nèi)部進行基于三角網(wǎng)格的非均勻縮放變換,而對剩余部分根據(jù)敏感區(qū)域的縮放變化做基于魚眼變換的壓縮.實驗結(jié)果表明,圖像敏感區(qū)域內(nèi)部失真很小,實驗結(jié)果很好地保持了敏感目標的特征,縮放效果較傳統(tǒng)內(nèi)容感知圖像縮放方法有明顯改善.
[Abstract]:In order to solve the uncertainty of the scaling effect of traditional content-aware image scaling technology, and the method is not suitable for multi-spectral remote sensing images, a new image scaling method combining triangular mesh and fish-eye transform is proposed. The sensitive target area is marked according to the feature recognition of the sensitive target, and then the non-uniform scaling transformation based on triangular mesh is carried out inside the labeled sensitive area. The remaining part is compressed based on the fish-eye transform according to the zoom change of the sensitive region. The experimental results show that the internal distortion of the sensitive region of the image is very small, and the experimental results keep the characteristics of the sensitive target very well. The scaling effect is obviously improved compared with the traditional content-aware image scaling method.
【作者單位】： 北方工業(yè)大學計算機學院;北方工業(yè)大學電子信息工程學院;
【基金】：國家自然科學基金項目(61371143) 北方工業(yè)大學基于內(nèi)容感知的最優(yōu)圖像縮放技術(shù)研究與應用科研平臺(XN054);北方工業(yè)大學優(yōu)勢學科項目(XN078)
【分類號】：TP751

【相似文獻】

中國期刊全文數(shù)據(jù)庫 前1條

1 吳均,朱重光;帶銳度保持的斜投影B-樣條圖像縮放[J];遙感學報;2002年02期

中國碩士學位論文全文數(shù)據(jù)庫 前1條

1 邢闊;基于內(nèi)容感知的遙感圖像縮放技術(shù)研究與實現(xiàn)[D];北方工業(yè)大學;2017年

，

本文編號：1544816