| Many insect are very small and most of the migratory behavior occur at night,thus it is difficult to see with the naked eye and observing the insect need specialized monitoring equipments.Water exist in the body of insects and the radar waves can be received again by the insect reflection,so by analyzing the azimuth distance height,speed flying insects and other information available radar echo,a new observation of insect migration route can be found and insect radar for monitoring migratory insects is designed.The main ecological areas of most the insect population are in the low altitude range,and there are few reports that studied the migratory pests in the lower air with radar monitoring.The scanning radar or vertical radar was generally used to monitor the insect both at home and abroad,the bandwidth of which was a little narrow,and could not need the requirements,so it needs a more greater bandwidth.This paper proposed a new idea about ultra-wideband radar monitoring insect insects and using a ring of log-periodic antenna array,which could meet the needs of different sizes to monitor the insects.Under the support of the National Natural Science Foundation of "Study of Active and Passive Location for Detecting the Flying Insect of Low Altitude Based on Ultra Wide Band Radar" fund and the graduate innovation special fund "Rice pests intelligent identification and classification system",our work mainly focused on studying the ultra-wideband radar for insects and employing the Doppler principle,MUSIC algorithms and other modern spatial spectrum estimation technology to get the target insect migration,speed and direction.The ground clutter interference suppression problems were studied by combining the key technologies of wavelet analysis and principal component analysis.By carrying out the ultra-wideband radar signal processing and simulation experiments,and introducing the harmonic radar technology to suppress the interference of metal and electronics products,it was showed that the noise ratio at a higher state,a higher angle resolution appeared in circular antenna array,which improved the estimation algorithm of broadband DOA was helpful to improve operational efficiency.Comparing the different sizes of MUSIC spectrum for insects,it was found that theoretical results were same.As the enemy of crop pests,to accurately predict pest is an important prerequisite for the forecast and control pests.The manual identification is time-consuming,and most determination is confined to a particular section because of the great varieties of insects.However,due to technical limitations of the radar itself,it could not reach the requirement to identify the target species and must need some other aids.Our paper used the digital image processing techniques to identify the principle objective.By using the computer and network technologies to analyze the insect image,taking rice pests for example,it was proposed that a discrete sine transform cooperated with the principal component analysis module and a method was designed that used computer software systems to realize the insect automatic image recognition.Experiments showed that the module 2DPCA intelligent image recognition method had the ability to identify.This method was used to compare with the conventional linear identification methods(such as PCA method),it had a little advantage in the acquisition value of the identifying characteristics of speed and this method was more simple,and its robustness characteristic recognition result was more evident. |
PDF Full Text Request