Research On The Key Technology For Smart Security Village House

Urban villages are a product of the urbanization process.The villages have complex environments,backward facilities,large floating populations,and frequent violations.However,the current supervision of violations in urban villages generally relies on manual management,with a large workload and low effectiveness.One of the core functions in the field of "intelligent security" is to detect or identify violations,persons and vehicles in violations.It has become an effective means to solve the current security problems in urban villages,and will surely become an important part of the construction and development of the village in the future.This paper takes the village in the city as the background and summarizes the previous research results,and conducts in-depth research on the key technologies of the detection of violations in the field of intelligence security,involving the detection and identification of the violations themselves,their main actors and vehicles.And implemented a deep learning-based violation detection and identification system to provide a feasible solution to the problem of frequent violations in urban villages that are difficult to manage.The specific work is as follows:(1)On the basis of constructing a data set of violations,a research on the detection of violations based on YOLOv4 is carried out.Use a smartphone to take 725 photos of violations,and use Adobe Photoshop CC to preprocess the images.Use label Img to define5 target objects and 4 types of violations to construct a data set of violations.Then train the violation detection model based on YOLOv4 to identify 5 target objects,and uses Center Net and FSAF respectively for comparison experiments with this model.The experimental results show that the recall rates of YOLOv4 on cage,stall,hawker,pitchman,umbrella are respectively.The accuracy is 91.14%,84.75%,91.23%,91.41%,and 83.85%,the accuracy is 94.74%,86.96%,94.3%,88.64%,and 94.13%,the average accuracy is95.44%,86.41%,93.98%,92.78%,and 89.88%,F1-Score is 93%,86%,93%,90%,and89%,m AP is 91.7%.YOLOv4 is better than Center Net and FSAF in the detection of violations,and can effectively detect violations.(2)Research on the identification model of violators based on LSCNN.In order to solve the problem that when a violation occurs,the violator escapes after encountering the law enforcement and cannot trace his identity.This paper trains the LSCNN violator identification model to realize the identification and appraisal of the person who violated the rules,and compare the performance of the model with the random forest model,the results show that the F1 value of the model on the test set is 98.60%.The experiment shows that the performance is better,and it can effectively identify the offender and capture the identity information of the offender.(3)On the basis of constructing the licence numberr data set,carry out the research on the identification technology of violating vehicles based on DFVGG.According to the characteristics of Chinese license plates,this paper constructs a licence number data set containing 26000 character pictures.Aiming at the problem of traffic congestion caused by violators using cars to occupy roads and parking illegally,this paper trains the DFVGG vehicle recognition model to realize the capture and identification of vehicles that violate the rules,at the same time,a performance comparison study between the model and the random forest model is carried out.The F1 value of this model on the licence number test set is 97.50%,and the experimental results are better than the random forest model.(4)Designed and implemented a violation detection and identification system.Comprehensive application of the above research results,a violation detection and identification system based on the SSM framework is constructed,which can realize the functions of detecting and querying violation records of violations in urban villages,their actors and vehicles.Based on deep learning,this paper conducts research on the detection technology of violations in urban villages,their main actors and vehicles,and builds a detection and recognition system for violations,which provides a new way for urban villages to solve the problem of frequent violations and promotes the development of urban villages.Intelligent security construction provides solutions with reference value.