透過您的圖書館登入
IP:18.224.44.108
  • 學位論文

可適性亮度修正於視訊應用之演算法及架構設計

Algorithm and Architecture Design of Adaptive Brightness Correction for Video Application

指導教授 : 曹恆偉
共同指導教授 : 范育成(Yu-Cheng Fan)

摘要


由於數位成像技術的發展,使人眼視覺得以儲存及處理,在一般消費性電子上,現行數位攝影應用早已變得相當熱門,例如專業的單眼相機及內建於行動電話的攝影功能;當我們執行數位成像時,一些環境的因素及攝影過程的行為有可能影響拍攝結果影像的品質,例如機體的晃動和周遭環境的亮度改變;當我們在暗夜或者低亮度的條件下拍攝,成像結果的可見度會因光線反射不足而降低,因為光源的分佈亦決定了物體表面的光反射量,所以不均勻的光源即可造成拍攝結果影像失去原本期望的特徵。   直方圖等化(Histogram Equalization)是最常使用的影像強化技術,其可有效地以全影像的機率分佈密度來重整原始直方圖,進而增加影像的可見度,因為區域影像的亮度分佈範圍不會和全域亮度分佈範圍相同,在硬體實現時須注意到計算的有效性,分析目前的方法發現研究的重心應放置在亮度強化幅度的可控制性及計算複雜度的降低。   我們提出區域雙直方圖等化,其實現操作可分為三個部份:亮度評估,雙線性等化和可適性亮度限制。為了分析區域亮度的分佈,我們使用亮度估計去實現,其中會搜尋亮度相似的像素位置,再以加權方式增加相似像素的影響,如此作法雖可有效地的分析區域亮度變化,但計算複雜度亦相對地增加,在區域亮度強化的處理中,其結果可能會受周遭陡然變化的區域像素干擾,在此應用加權計算即來消除不相似亮度值的干擾,可增加相似的影像特徵值,區域雙直方圖等化不再以全域平均值控制區域亮度強化的幅度,改以上述的加權平均值操作,接著以雙線性等化的計算處理,將原始輸入亮度值轉換,雙線性等化的計算用以取代傳統直方圖等化的計算,由於影像可能含許多區域亮度需調整,所以提出可適性亮度限制曲線來簡化區域影像的修正,利用單一曲線輔助修正控制。目前可適性亮度修正技術未有合適的架構設計,再加上可適性亮度修正的架構設計存在許多問題,如計算複雜度及資料流之排序週期,針對複雜度的問題,我們提出取樣式處理來減少外部記憶體的資料搬動,區域雙直方圖等化的最大的特徵是其較符合架構設計的須求。   以標準的晶片實現流程來實現我們的架構,當我們完成系統層的模擬,接著便可以硬體描述語言描述架構設計,在完成其模擬及結果的比較後,我們便可開始以電腦輔助設計工具來完成實體層的晶片實現,合成時須以9.25 ns為週期限制來合成,其合成後邏輯閘數為69.5K,我們晶片使用UMC 0.18um 1P6M CMOS技術來實現,其晶片核心的尺吋大小為1.96mm X1.96mm,經由比較的結果,可以了解我們的晶片可以較少的週期實現一個移動向量的搜尋,以及晶片所使用的記憶體位元數及面積皆是合理的大小,其結果亦說明此晶片可以實現更大尺吋的影像或視訊應用,並且符合開始晶片的規劃,如此我們的晶片便可以有效地在MPEG系統中實現可適性亮度估計,並相容於目前視訊應用。

並列摘要


Due to the developments of digital imaging techniques, the visions of human eye become storable and removable. When using digital camera, few environment factor and user behaviors probably affect the quality of captured image, for example the camera motion and surrounding brightness. When we take picture under the condition of dark eve or low illumination, image visibility will be decreased by insufficient light reflection. Because the distribution of light source decides the reflection direction, unequal intensity may cause the resultant image lose expected feature. According to the overviews and analysis of adaptive brightness correction, we can understand that the techniques can provide an available solution for the interference of environment brightness in digital camera, but current methods are still no suitable algorithm design. Therefore, we propose the local bi-histogram equalization (LBHE) to remedy the local visibility and control the degree of enhancement. The procedure can be divided into four calculations: brightness estimation, bi-linear equalization, adaptive brightness constraint. We also present the sample-based data flow to reduce the data movement of external memory. We design hybrid architecture to implement adaptive brightness correction with MPEG system. Based on the specification of video application, our chip would be implemented with the real time requirements of 30 frame per second (fps) and D1 format (720x 480). The standard cell-based VLSI design flow is adopted to realize our work. Our design is synthesized with the timing constraint of 108 MHz and the total logic gate count is about 69.5K. The chip is implemented with UMC 0.18 µm 1P6M CMOS technology. The layout result present the core size is about 1.96 mm × 1.96 mm. Through the comparison content, we can find that our design can exactly satisfy the chip specification with reasonable implementation constraints.

參考文獻


[2] Y.-T. Kim, “Contrast enhancement using brightness preserving bi-histogram equalization,” IEEE Transactions on Consumer Electronics, vol. 43, no. 1, Feb.
[3] Soong-Der Chen, Adb. Rahman Ramli, ”Contrast Enhancement using Recursive Mean-Separate Histogram Equalization for Scalable Brightness Preservation,” IEEE Transactions on Consumer Electronics, vol. 49, no. 4, Nov 2003
[4] Yafei Tian, Qingtao Wan, Fengjun We, “Local Histogram Equalization Based on the Minimum Brightness Error,” IEEE International Conference on Imaging Graphics, ICIG 2007
[5] Chao Wang, Zhongfu Ye, “Brightness Preseving Histogram Equalization with Maximum Entropy,” IEEE Trans. On CE, Vol. 51, No.4, Nov. 2005
[6] Abduallah M. Alsuwailem, “A New Approach For Real- Time Histogram Equalization Using FPGA,” IEEE International Symposium on Intelligent Signal Processing and Communication System, IEEE ISPACS, 2005

延伸閱讀