Design And Implementation Of Remote Sensing Image Intelligent Interpretation System

Drived by the rapid development of science and technology,remote sensing images appear in a large number of people’s vision,and gradually become an important source of spatial geographic information data.Remote sensing image intelligent interpretation technology has more and more development space with the improvement of remote sensing technology.The current intelligent interpretation algorithms of remote sensing images have the problems of low accuracy and slow interpretation speed,and the existing algorithms do not make sufficient use of remote sensing image information.In this paper,the relevant basis and algorithm of remote sensing image intelligent interpretation are introduced,and the image semantic segmentation method which has better effect in visible light is introduced into remote sensing image intelligent interpretation,and the algorithm is improved.To realize the improvement of the interpretation speed and the reduction of the number of parameters of the intelligent interpretation algorithm of remote sensing images.At the same time,the intelligent interpretation system of remote sensing image is realized.The specific work of this paper includes the following four points:(1)A remote sensing image intelligent interpretation method based on Deeplabv3 is proposed and its performance is compared with other remote sensing image intelligent interpretation algorithms.The preprocessing method of remote sensing image data set is improved and the multispectral fusion method is introduced to increase the utilization rate of remote sensing image information.(2)To improve the proposed DeeplabV3 interpretation method.The interpretation accuracy is slightly improved by the improved method.The experimental results show that the proposed method has some problems,such as low interpreting speed and large number of parameters.(3)The MobileV2 method was introduced to improve the low speed of the interpretation method and the large number of parameters,which reduced the interpretation time by nearly three times and greatly reduced the number of network parameters.BCE and DICE loss combined with loss function method is adopted to improve the non-uniformity of samples in the interpretation data set.(4)Realization of remote sensing image intelligent interpretation system.Analyze the requirements of remote sensing image intelligent interpretation system and design the system as a whole,each module and database;Through the development of the system,combining the algorithm proposed in this paper and the previous algorithm to achieve interpretation management,image management,image preprocessing,interpretation results analysis and user management modules,the remote sensing image interpretation system is realized.