唐亞平+陳蘇婷
摘要:針對Harris角點檢測過程中存在定位粗糙、檢測精度不高以及檢測效率慢等原因,該文在Harris算法的基礎(chǔ)上,結(jié)合Harris算子和Forstner算子提出一種改進的亞像素角點提取算法。該算法采用一種逐層檢測策略,首先利用Harris算法進行角點粗略定位,首先對角點做一個初始選擇,利用圖像領(lǐng)域灰度相似度得到大部分角點的粗定位值,大大降低了算法的運算量,然后通過計算自相關(guān)矩陣的兩個特征值,利用特征值和閾值比較篩選得到全部角點的粗定位值,避免了CRF(corner reference function角點響應(yīng)函數(shù))的計算,最后利用Forstner算子對粗定位后的角點進行亞像素級精確定位。實驗證明,該算法不僅保證Harris算法的靈活性和Forstner算子的亞像素級精度,而且速度快,并且抗噪聲性能較強。
關(guān)鍵詞:Harris;Forstner;CRF;亞像素 效率;精度
中圖分類號: TU198+.3 文獻標識碼:A 文章編號:1009-3044(2014)19-4552-04
An Improved Subpixel Corner Extraction Algorithm
TANG Ya-ping, CHEN Su-ting
(Jiangsu Key Laboratory of Meteorological Observation and Information Processing, Nanjing University of Information Science &Technology, Nanjing 210044, China)
Abstract: In view of the rough positioning、inaccurate detecting and the slow efficiency in the Harris corner detection,According to this,an improved Harris algorithm is presented in this paper, on the basis of the combining with Harris algorithm and Forstner operator . The algorithm adopts a layered detection strategy, firstly using Harris algorithm for corner rough localization, considering the slow detecting, before calculating the CRF we made an initial choice to avoid large amount of multiplication, after geting the coarse position of angular point, we then calculate Forstner operator for accurate positioning. Experiments show that the algorithm not only guarantees the flexibility of Harris algorithm and the subpixel accuracy of Forstner operator ,but also improve the speed and have strong anti-noise performance. The algorithm overcomes the problem on the detection speed and precision in the Harris algorithm.
Key words: Harris; Forstner; CRF; subpixel; accuracy; detection efficiency
圖像特征點的提取是計算機視覺領(lǐng)域中非常重要的一步,其中,圖像角點是最常見的一類特征之一,它具有旋轉(zhuǎn)不變性、不隨光照變化等優(yōu)點,因此角點特征被廣泛應(yīng)用于物體跟蹤、三維建模、攝影測量自動化以及遙感圖像匹配等領(lǐng)域。
本文通過分析Harris角點檢測過程中的一些弊端,如在對角點進行選擇定位的過程中采用了大量乘法運算,運算量大,大大降低了算法檢測的效率,同時Harris算法只能檢測到像素級水平的坐標,對于精度要求高,需要準確定位像素坐標的場合不能夠滿足精度要求,該文在Harris算法基礎(chǔ)上提出了一種改進……