The Design Of Blind Instrument Based On Embedded Image Processing And Ultrasonic Ranging

According to the statistics of the World Health Organization,the number of blind people in the world is about 36 million,and the trend is increasing year by year.The number of blind people in our country has always ranked first in the world.Compared with normal people,the blind who lose their visual perception ability cannot obtain obstacle information and sign information in the environment,their daily life and travel become extremely difficult.Therefore,this paper studies and designs an intelligent wearable blind guider,which can effectively guide the blind to avoid obstacles,identify important traffic signs such as zebra crossings,and sidewalk traffic lights,as well as blind roads,the blind guider gives corresponding action instructions through voice.At the same time,the device uses ultrasonic ranging technology to solve the misjudgment of information caused by uncontrollable factors such as vehicles and pedestrians,to assists the blind guider for more accurate image processing.The blind guider adopts digital image processing technology to process the image in real time.The ultra-small embedded development board Nano Pi NEO2 is also used as the main control unit,aim to reduce the size of the blind guider to improve portability.The main work of this paper also includes: designing the hardware platform for the blind guider,preprocessing the image through operations such as cropping,median filtering,grayscale and perspective transformation,and using morphological edge enhancement to reconstruct and identify the zebra crossing area based on the stripe characteristics.In order to solve the problem that the traffic light is difficult to identify under a large background,the Adaboost classifier algorithm is adopted to accurately position the traffic lights,and a "column pixel judgment method" is proposed to accurately distinguish the traffic light's color under the HSV color model.The watershed algorithm is used to segment the blind road area through its texture and edge features.When the recognition of the zebra crossing and blind road track by the blind guider is interrupted,the function of ultrasonic detection will assist the guider device to eliminate the interference of the obstacle information on the image recognition,the device reminds the user to avoid the obstacle but the image detection result will not fed back.After the object disappears,the blind guider continue to carry out relevant image detection and recognition of the current environment,potential dangers caused by uncontrollable obstacles such as vehicles and pedestrians will be avoided.The blind guider uses the Open CV vision library to complete part of the image processing algorithm,the information obtained by the blind guider is also fed back to the user through voice prompts by TTS speech synthesis software.Experiments tests have shown that the blind guider can effectively guide the blind to travel safely,assist them to avoid obstacles,and recognize important traffic signs such as zebra crossings,traffic lights,blind roads,and give voice commands such as waiting,stopping,and passing.The frame rate of the blind guider recognition under the traffic light recognition is 11 frames per second,which achieves the effect of real-time detection.Ultrasound-assisted detection effectively solves the misjudgment of the information of the blind guider caused by uncontrollable obstacles such as vehicles and pedestrians,and increases the safety of the instrument.The blind guider collects the image through the camera and processes the image information into voice information through the embedded microprocessor,the blind can quickly understand and react,choose the correct route of travel,makes travel become safe and convenient.