Development Of Bionic System Based On Auditory And Vision Joint Target Tracking

With the development of computer technologies,electronic sciences and network technologies,intelligent monitoring has received more and more attention.The traditional video surveillance methods can meet the basic needs of security defenses under normal circumstances,but the inadequacies are increasingly evident.First of all,the traditional methods cannot actively track the surveillance target,but can only passively monitor the target of the predetermined angle.Secondly,the traditional methods cannot respond to the abnormal situation outside the viewport,there is a monitoring blind spot.There is a monitoring blind spot.Finally,the targets in the viewing zone cannot be identified,so that they can't be focused and centered to improve clarity.In response to these shortcomings,combined with the development of sound source location technology and image processing technology,we combined sound source location technology and image processing technology,used bionic principles,and utilized sound signal characteristics to design a more effective monitoring system.The innovations and main work of this article are summarized as follows:First,at the auditory level,a digital omnidirectional MEMS microphone is used as an array sensor to design a new sound source localization system.Unlike traditional electret microphones,digital MEMS microphones have the characteristics of high sensitivity,flat frequency response,miniaturization,low power consumption,good omnidirectionality and consistency,and the digital output facilitates system integration.We put the FPGA data acquisition module,DSP data processing module and MEMS microphone array on a circular PCB board with a diameter of 9cm.We have achieved high system integration,portability,real-time performance and low power consumption.The audio endpoint detection algorithm is used to preprocess the signals collected by the array,effectively reducing the amount of computation.A frequency?domain beamforming algorithm for wideband was used to effectively implement the positioning of far-field sound source signals.The sound signal with non-stationary features was tested and verified in a non-consumed acoustic room environment.The verification results show that the positioning accuracy is high and the real-time dynamic effect is good.Secondly,at the visual level,we briefly describes the image processing methods used in this paper,introduces several target detection methods and analyzes their advantages and disadvantages.We proposed a method of using the method of detection to detect the target and using MeanShift algorithm to track the target.The tracking success rate of the visual tracking module was tested in the outdoor environment with relatively complex indoor environment and background.The test results proved that the tracking success rate of the system is high and the real-time performance is good.In the end,we use matlab and MFC to design and develop the sound source detection and location system in the monitoring system.