Software Design And Development Of Handheld Magnifier System Based On Linux

Because of the lack of vision,low vision people cannot read or distinguish objects fastly like normal people.The single-handheld magnifier system show a convenient and efficient way to help low vision people in reading and distinguishing.At present,China have done various of research in the field of electronic magnifier,a variety of visual aids products emerge in an endless stream,and the functions of hand-held electronic magnifier become more and more powerful.This thesis aims to design a handheld magnifier system which is cost-effective.This system designs the software framework based on Linux system.In order to enhance the user experience,system is combined with various of peripheral and intelligent algorithm framework and hardware acceleration platform of Hisilicon.The main contents of this thesis are as follows:(1)The overall design of the hand-held magnifer system.Determine overall framework structure according to the system requirement in actual use case,thie part alse design the hardware scheme and software architecture.Besides,design an automatic upgrade function which is independent of the software framework.peripheral driver migration is also mentioned in this part.(2)Software design of video capture,process and display.Combined with the hardware and software characteristics of Hisilicon platform,the scaling and color transforming functions of traditional manifier are optimized.According to the system power consumption and the user’s demand,the camera adaptive switch and screen adaptive rotate function are designed and implemented,and the anti shake algorithm in the distance camera is optimized by combining with gyroscope data.(3)Qt based user interface software design.Combined with the needs of low vision people and the rich class encapsulation mechanism of Qt,a set of functional modular human-computer interaction architecture is designed,which effectively handles the various input operations of users.Aiming at the problem of screen cutting caused by the poor rendering performance of Qt software,a drawing method is designed to give full play to the hardware characteristics of Hisilicon,and a special drawing plug-in for Hisilicon is designed in combination with Qt characteristics.(4)Intelligent processing algorithm design and software development.Aiming at the problem that low vision people are not sensitive to the edge of objects,the edge enhancement algorithm is optimized by using the hardware characteristics of ive.An object recognition algorithm based on CNN+SVM is proposed,training data set,network and classifier in Caffe deep learning framework,optimizing the feature extraction process and classification recognition process by combining the ive characteristics,together wirh the speech function,low vision people can effectively recognize objects.In order to reduce the power consumption of the device,an automatic shutdown function based on mobile detection is designed.