Video Gesture Recognition Smart Home Control System Based On FPGA

With the development of the national economy and the improvement of the level of science and technology,artificial intelligence from the theory gradually into the real life,and human-computer interaction as an important branch of the field of artificial intelligence,also derived more and more interactive ways,the demand of human-computer interaction for the mature gesture recognition system is further enhanced.Compared with using devices such as mouse,keyboard and touch screen to control smart home,the gesture-based smart home control method has the characteristics of simple operation and contactless control,which has important research significance for the existing smart home control methods.This thesis has done research on the realization of gesture recognition smart home system on the embedded hardware platform.Using Field Programmable Gate Array（FPGA）as the hardware platform,a gesture recognition system is designed,which can realize real-time gesture recognition and control of smart home devices.In the algorithm design,first preprocess the collected gesture images,including color space conversion and Gaussian filtering.At the same time,a fast and effective gesture segmentation algorithm is proposed,use watershed segmentation and wrist segmentation algorithms to obtain independent palm images;Extract features such as the area,perimeter,and Fourier descriptor of the segmented gesture image through a feature extraction algorithm.According to the area and circumference ratio feature of the palm,compare the gestures in the sample library to narrow the search range,and finally use the Fourier descriptor to perform Fourier correlation analysis to identify the final gesture.On FPGA functional design,the system adopt top-down design ideas,uses Verilog hardware description language to program each part of the function,and divides the various modules of the system according to the function.First,initialize the camera configuration,obtain the palm and arm area through the image acquisition and preprocessing module,then drive the SDRAM storage module to cache the video stream data and process it across clock domains,obtain gesture instructions from the output data through the gesture recognition algorithm module.Finally,data transmission is carried out through the Wi Fi communication module to complete the control of smart home equipment.The test results show that the FPGA-based video gesture recognition smart home control system designed in this thesis can recognize a total of 10 gestures,with a recognition rate of 96.75%,frame processing speed reaches 24.39 fps.At the same time,the control of smart home equipment is tested,the average success rate of several kinds of furniture equipment control is over 96%,and the average response time is only 0.458 s.which can meet the actual use effect.