The Research Of Feature Description And Video Content Detection About Power Meter Equipment

With the rapid development of science and technology,intelligent monitoring applications have been into people's lives,like the family security,and eye system,and have taken great changes for us.The monitoring system represents part of the intelligent life and it is used in all aspects of life.At the same time,the demand for intelligent monitoring is increasing with the rapid development of the power system.There are two existing types methods of monitoring in power system: fixed equipment monitoring and inspection robots.Fixed equipment monitoring commonly stated in the substation or other fixed place,mainly located in the substation on both sides of the road and equipment,and are used for timing monitoring the environment and equipment.As for inspection robots,they have the fixed routes,and commonly used in infrared temperature detection and height of the wire.However,in these two methods,most of the photos and videos that need to be reviewed manually,so that they cannot get fault information in real time.They are often used to look back.At the same time,the large amounts of data produced by these two methods have also brought great difficulties to computer hardware and storage.As a result,content based video retrieval(CBVDS,Content Based Video Database)has gradually became a new trend of video image monitoring and detection.Therefore,based on the investigation of the power system equipment,equipment surrounding environment and equipment detection requirements,we set up the feature description and video content detection platform for power meter equipment.First of all,the improvement of meter equipment detection algorithm to detect gauge and high temperature failure of power meter equipment.Next,the creation of classification about some power equipment,including the classification of equipment and common equipment failure.Finally,the establishment of a content based feature.And in order to reduce the occupied space of the video,the MPEG-7 which is a standard about a description of the multimedia information,are used to descript and storage the equipment uniformly.The main contents of this paper are as follows:1)Through the analysis of some common faults of some power equipment and the demand of power system to realize automatic equipment detection,we found that there was a lack of unified description method for power equipment.Therefore,we introduced the MPEG-7 description specification,and presented the power equipment classification framework based on the MPEG-7 standard.The main contents of the classification framework included: the classification of the power system equipment,the summary of the common faults of the equipment,the different detection algorithms.According to the framework and MPEG-7 specification,a unified description of the power equipment could be carried out,and the framework could provide the basis for construction of the feature database.2)By analyzing the fault of power meter equipment,like ultra-high on the temperature,oil level gauge etc.,we studied the Hough transformation detection algorithm for meter equipment.We found that the algorithm was fast but the false detection rate was high.So in order to realize the accurate detection of the status of electric meters and meters,we improved Hough transform and presented Hough transform based on standard Euclidean distance and threshold.At the same time,we found that moving targets such as birds and pedestrians around electrical equipment may cause equipment to enter an insecure state.So we used Mean-shift algorithm for tracking and detection.However,the algorithm had poor effect of original algorithm on occlusion and fast moving target tracking.Therefore,we used the moving target detection method of Mean-Shift by Gabor-LBP and the color feature fusion based on prediction.3)Based on the analysis of feature extraction in pointer meter equipment and mobile target detection,combined with the actual situation of power equipment,we designed a feature database that could uniformly store feature and describe the hierarchical characteristics of devices according to the description framework established in the early stage.For a part of the database,it could storage the low-level features of devices,like color feature,color histogram,contour feature,edge feature etc.;the other part was used for the semantic information storage,for example,device number,name,location and other equipment information to represent pictures,video and other multimedia information.The feature storage was stored in the XML language format in order to reduce the amount of space occupied by the direct storage of multimedia information.In this study,based on the demands about the electric power system,electric power equipment and the classification of common faults,combined with different types of detection algorithm and feature extraction algorithm,we built the power device description and video content detection platform based on MPEG-7 with the database.It mainly realized the functions of user management,table meter detection,database query and so on.The platform provides a certain basis for the future development of power equipment monitoring system.