使用雙線性內插法之嵌入式系統影像縮放技術研究

Study on Image Resizing for Embedded Systems by Using the Bilinear Interpolation Method

李秉儒(Ping-Ju Li)

指導教授 ： 曾傳蘆 ； 王順源

國立臺北科技大學/電資學院/電機工程系研究所/碩士(2010年)

摘要

在嵌入式系統中由於顯示器像素量大多是固定的尺寸，且其解析度通常不會很高，所以在交換不固定大小的影像檔案或視訊資料時，會因為像素的不同造成顯示上的困擾。
針對上述問題，本論文設計一影像縮放介面，可以適當地調整影像尺寸的大小，以符合裝置本身的設定與提高顯示器顯示的效果，使圖片失真效果降低，且增加使用者觀看之舒適性。所採用的原理為雙線性內插法，依照欲補償像素與相鄰四點的實際像素距離的遠近的比例求得補償值，使得相鄰像素之間更有連續性。所使用的嵌入式系統為德州儀器(Texas Instruments)針對影像處理推出的數位訊號處理器(Digital Signal Processors，DSP) Davinci系列，以及賽靈思(Xilinx)現場可編程輯閘陣列(Field-Programmable Gate Arrays，FPGA)，再搭配友達(AUO)薄膜電晶體液晶顯示器(TFT-LCD)，及一個簡單的鍵盤。作業系統以Linux為核心，可運行應用程式，再以C語言撰寫應用程式，將讀入的圖片以雙線性原理調整至顯示器可顯示的範圍，並提供選取框可透過按鍵決定選取範圍後再進行放大。為提高銳利度，介面提供高通濾波功能來銳化圖片，藉以改善使用雙線性內插法放大而產生的模糊影像。
由所發展的嵌入式系統圖形介面操作結果可知，本論文所提方法確實具有實用性。除此之外，本論文利用訊號峰雜訊比(Peak Signal-to-Noise Ratio，PSNR)值做為性能指標並進行實驗，並使用Matlab的Simulink求出經過高通濾波器的圖片與未經過的圖片個別的PSNR值並進行比較。由實驗結果可得知經過高通濾波器的圖片呈現效果較好。

關鍵字

嵌入式系統 ； 雙線性內插法 ； 影像縮放

並列摘要

In embedded systems, the number of pixels of LCD display is fixed, and the resolution is usually not very high. Therefore, as exchanging the unfixed-sized image files or video data, the unmatched image pixels often cause display problem.
To solve the above problem, this thesis designs a video resizing interface to adjust the image size in order to match the device setting and improve the display performance of the monitor. Moreover, the proposed method can reduce the image distortion, and make the user more comfortable while watching the screen. The resizing algorithm adopts the bilinear interpolation method. This method calculates the compensation pixels according to the distances between considered pixel and the adjacent four pixels and thus makes the values of pixels smoother between adjacent points.
The embedded system used in this thesis includes the Davinci-series digital signal processor, which is released by Texas Instruments for image processing, Xilinx-Series field programmable gate array, AUO thin TFT-LCD, and a simple keyboard. The operating system is Linux, which can run applications. The bilinear interpolation method is programmed in C language to adjust the picture size, so that the picture can be fully shown on the screen. Moreover, the designed user interface provides a marquee which selects the portion of image to zoom in. In addition, to improve the sharpness, the user interface has a high-pass filter to sharpen the blurred image enlarged by the bilinear interpolation method.
From the results of operating the developed graphic user interface in the embedded system, it is found that the proposed method is feasible and practical. In addition, the Peak Signal-to-Noise Ratio (PSNR) value is adopted as the performance index and several experiments are made to compare the PSNR by using Matlab/Simulink software. The experimental results show that the PSNR values of the pictures processed by the high-pass filter outperform that of the pictures without using the high-pass filter.

並列關鍵字

Embedded Systems ； bilinear interpolation ； image scaling

參考文獻

[1] J. D. Ouwerkerk, “Image super-resolution survey,” Image and Vision Computing, vol. 24, 2006, pp.1039-1052.

[2] N. Shezaf et al, “Adaptive low complexity algorithm for image zooming at fractional scaling ratio,” Proceedings of the 21st IEEE Convention of the Electrical and Electronic Engineers, April 2000, pp. 253-256.

[4] M. C. Tsai and P. Y. Huang, “Design of scan converter using the locally 2-D bilinear interpolation” , 2006 IEEE International Conference on Systems, Man, and Cybernetics, October 8-11, 2006, Taipei, Taiwan, pp. 3961-3966.

[7] M. A. Nuno-Maganda and M. O. Arias-Estrada, “Real-time FPGA-based architecture for bicubic interpolation: An application for digital image scaling,” International Conference on Reconfigurable Computing and FPGAs, 2005, pp. 1-2.

[9] R. Wang and S. Yang, “The design of a rapid prototype platform for ARM based embedded system,” IEEE Transactions on Consumer Electronics, vol. 50, 2004, pp. 746-751.