Research On Automatic Analysis Of Image Intelligence In Big Data Environment

With the development of information technology,especially the artificial intelligence and cloud computing,the intelligence analysis technology under the big data environment has become a hot research field,and the information analysis technology under the big data environment has gradually become an important support for battlefield situation analysis and operational command decision,thus the modern warfare has become a typical “intelligence processing” war.With the rapid growth of the amount of image intelligence data and the diversified of image intelligence sources,the study of image intelligence automatic analysis will help to obtain the important information of image intelligence quickly and accurately.In order to meet the specific task requirements of target recognition,image mosaic,people detection in the automatic analysis under the big data environment,besides to increase the accuracy and speed of image recognition,we launches a targeted study on the key technology of intelligence rapid retrieval and recognition,as well as the battlefield analysis method based on automatic analysis technology of image intelligence,on the basis of the key technology of target recognition and image mosaic in feature matching,the method of object detection and the big data analysis.The specific research contents are as follows:(1)Firstly,SIFT and SURF feature extraction methods are studied,and a weighted template library method is proposed to greatly improve the accuracy of image recognition based on feature point registration.(2)To study and compare the performance of SIFT and SURF algorithms,two or more images based on feature point registration technology and video sequential frame stitching technology.(3)Analyzing the commonly used skin color detection and Ada Boost algorithm in face detection,two new face detection methods are proposed.The experimental results show that both new methods greatly reduce the false alarm rate.(4)Analyzing the task requirements,technical requirements and data characteristics of battlefield reconnaissance image information support,putting forward the automatic analysis flow of battlefield image information and video data respectively,putting forward the overload learning machine cluster algorithm to meet the needs of fast image information retrieval task,and putting forward the detection,extraction,classification,recognition,tracking and speed measurement methods of moving target image in video data.