在現今人機互動的系統中，觸控手勢和即時影像操作，如影像的平移、縮放及旋轉，已經漸漸成為必備的功能。而目前這些運算主要都是透過後端的中介軟體來實現，這將會加佔用更多的系統資源，並不適合要求即時且資源有限的嵌入式系統。本研究提出一個整合觸控與即時影像操作的人機互動介面控制器。對於即時的影像處理，使用CORID(COrdinate Rotation DIgital Computer)計算器來取代原本複雜的三角函數計算，並使用雙立方模板插補(bicubic spline)演算法進行影像的插補工作，再透過顯示控制器呈現在LCD上，最後整合智慧型即時手勢辨識控制器。我們以MIAT方法論將其演算法整合為一個全硬體化的人機互動介面控制器，於FPGA(Field Programmable Gate Array)平台進行硬體驗證，並使用管線化(pipeline)架構設計硬體電路來提高系統效能。實驗分析的結果顯示，影像操作的處理速度較軟體執行來的快，適用於嵌入式系統及可攜式電子產品。In today's human-machine interaction(HMI) systems, the touch gestures and the real-time image manipulation , such as image translation, scaling and rotating, have became the must-have features. These operations conventionally are implemented through the embeddedmiddleware of back-end system. This software's solution is not suitable for real-time requirements and limited resources of embedded systems because of consuming too much system resources, so in this study, we design a high efficient hardwired HMI controller integrating the image manipulation and touch gesture. We implement a CORDIC processor for the complex trigonometric calculations, and a bicubic spline accelerator for real-time image interpolation. Finally, we integrate the touch gesture controller with real-time gesturerecognition engine. The hardwired controller is realize by using MIAT embedded hardware design methodology and verified on a FPGA platform. In the experiments, the proposed controller provide the better performance of the image manipulation than the software’s solution, which is suitable for power-aware , low-cost embedded system and portable electronic products.
