|Abstract: ||本研究以現場可編程輯閘陣列 (Field Programmable Gate Array; FPGA)開發平台結合數位影像處理技術來設計盲多重障礙者溝通輔具，研究中將FPGA所設計之系統劃分為硬體模組與應用C語言軟體方式來進行軟硬體共同設計。在硬體模組方面，設計CMOS擷取模組、RAW2RGB (Raw to RGB)模組、膚色擷取模組、影像濾波模組、注音編碼模組、CMOS主控端模組及LTM主控端模組等七個模組的硬體電路。在軟體設計方面，依據SOPC (System on a Programmable Chip)技術在晶片系統內嵌Nios II軟核處理器，並以C語言來實現影像連通區域分割演算法、人臉膚色區塊中心位置計算與八方向判別，最後計算人臉膚色區塊中心點之移動座標作為八方向判別訊號，其訊號依照數碼拼音表定義經由注音編碼模組運算後傳回Nios II處理器做注音符號與聲調符號顯示。|
This study aimed to design an assistive communication device using the FPGA (Field Programmable Gate Array) development platform, combining digital image processing technology, to help visually impaired people with multiple disabilities. In this study, the system designed with FPGA was divided into hardware modules and applied C-language software for co-design of software and hardware. In the aspect of hardware, the hardware circuits for seven modules, including CMOS Capture Module, RAW2RGB Module, Skin Detection Module, Image Filter Module, Zhuyin Encoded Module, CMOS Master Module, and LTM Master Module, were developed. In the aspect of software design, with the SOPC (System on a Programmable Chip) technology, the Nios II soft processor was implemented in the system-on-chip. And C-language was applied to perform the image connecting area dividing algorithm, calculation of center of human facial skin color block, and 8-direction distinguishing. Finally, the movement coordinates of the centers of human facial skin color blocks were used as signals for 8-direction distinguishing. The signals were encoded using the Zhuyin Encoded Module based on the definitions from the digital phonetic symbol table and then sent back to the Nios II processor to present Zhuyin symbols and tone symbols.
This study applied digital image processing technology to develop and implement an assistive communication device through the method of software/hardware co-designing on a FPGA chip featuring low power, low cost, and fast production. Although the overall calculation speed of the system is slightly slow, the result of this study can still be used as a reference of image-based embedded assistive communication device for visually impaired people with multiple disabilities.