Design And Implementation Of Static Gesture Recognition System Based On FPGA

As the basic expressive pattern of human communication, gestures play a very important role in interpersonal communication. As same as speech language and written language, gesture language can be used to help express people’s thoughts and emotions. With the rapid development of computer science and digital image processing, the static gesture recognition System based on video is becoming an important trend. The system based on FPGA(Field Programmable Gate Array) can meet the needs of high speed image processing with its hardware features. Compared with the implementation on PC (Personal Computer), the gesture recognition system based on FPGA has more application prospects.This paper designs a static gesture recognition system based on FPGA. The system uses Altera DE2development board as the hardware platform. It can collect gesture image with the CMOS image sensor and display the environment and processed image with VGA in real time. The system can identify ten different static gestures and control the particular LEDs.The system consists of four modules:core-control processing module, gesture image capture module, gesture image monitoring module and identification result display module. The system is implemented mainly in hardware, and EP2C35F672C6is used as the core of the system. Core-control processing module implements the main algorithms of the system, and it can be divided into three parts:image processing, feature extraction and matching. Firstly, the image processing module adopts an algorithm based on skin color clustering. It converts the image into a binary one, and then uses filtering, expansion and edge detection. Secondly, feature extraction module uses Hu moments algorithm to extract gesture feature parameters with the perimeter and area of the gesture image. Finally, feature matching module uses the improved template matching algorithm to recognize the gesture secondarily, and judges whether the gesture is defined one. If it is the definition one, the corresponding LED goes on. Otherwise, the LED is off. After detailed design and implementation, integrate various modules into a complete system. Finally, after repeated debugging and improvement, it realizes the system functions on DE2. The experiment results show that the system has implemented the basic functions of the static gesture recognition system, and it has its own advantage on increasing treatment effect and accuracy.