Research And Implementation Of Four Rotor Drone Detection Technology Based On Sound Recognition

With the development of the urbanization,civil unmanned aerial vehicle(UAV)brings portability to people's daily lives,such as aerial photography,cruise,reconnaissance,fire control,but also includes security risks.Based on the current lack of regulation of the UAV,it is an urgent need to have a practical and low-cost UAV detection system.Sound is one of the main characteristics of the four rotor UAV,this paper studies the problem of the UAV sound recognition.Firstly,the UAV detection technology and sound recognition technology are summarized and compared to the current mainstream,analyze the advantages and disadvantages of various detection techniques,and according to the application background,this paper focus on the UAV detection technology research based on sound recognition technology and application in UAV detection system design.Taking sound as the significant feature of the UAV into account,the paper full analyze the UAV's sound characteristics,respectively.The sound characteristics of UAV's are extracted from different angles such as time domain and frequency domain,including the Mel-frequency Cepstral Coefficients and Linear Prediction Cepstrum Coefficient.The FISHER criterion method is applied to the extraction of the characteristic parameters.The characteristic parameters are selected and dimensionality reduced,and the features of UAVs are displayed with less dimension as much as possible.The support vector machine(SVM)method and the Gaussian mixture model(GMM)method are applied to the four-rotor unmanned aerial vehicle(UAV).The recognition performance of the two algorithms is simulated and analyzed by MATLAB.Experiments show that the FISHER criterion is used to identify the sound of the UAV with the support vector machine method,which has the advantages of high precision and fast speed.At the end of the paper,the software design and implementation of the UAV detection system based on sound recognition have been researched,including PC software interface design and recognition algorithm design,and the DaJiang ghost series UAV as an example to test the performance of the system.The system designed in this paper has the advantages of stability instantaneity,high recognition accuracy and practical value.