A Perceptual Image Calculation Method For Riot Control Vehicle Design Evaluatio

Anti riot vehicles,a special vehicle to quell group riots and combat terrorist activities,play an important role in the implementation of complex and changeable peacekeeping tasks in the increasingly severe international security situation.Because of intensive competition of market,the appearance design has become an important standard to evaluate the quality of riot vehicles.In addition,due to the development delay in our country in the field of riot vehicles,technology still fell behind with some western countries like United States,Germany,as well as Russia.Therefore,it would be necessary to draw a set of complicated evaluation standards,which is of great significance to the anti riot vehicles’ boost in our country.Consequently,the aim for this paper is to construct a more objective evaluation system to assist designers’ work on the design of riot vehicles with scientific methodologies.Firstly,we collect massive samples and data information with the web crawler in Python.Through analyses of these samples and data with morphological analysis method,we encode,characterized the morphological characteristic parameters,and the typical sample database of riot vehicles is obtained through cluster analysis method.Secondly,we collect the perceptual semantic vocabularies of riot vehicles,and classified and merged them by KJ method.We quantitatively analyzed these combined perceptual semantic vocabularies with PCA(principal component analysis)to obtain typical representative of the design style of riot vehicles in the same dimension with no intersection.Then we introduce the Likert scale to evaluate the typical samples of anti riot vehicles to get the modeling feature parameters that are related to the image values.Then we eliminate the samples with similar shapes by cluster analysis,so as to achieve mapping relationship between typical shape samples and the perceptual image.We introduce the improved BP network to construct the predicted model with the mapping relationship.We verify the optimized algorithm has higher fitting degree by comparing it with the traditional BP neural network.We select the combat anti riot vehicle to build the evaluation model.The perceptual image of combat riot control vehicle is weighted by Stochastic Analytical Hierarchy Process(SAHP),and the evaluation results of each expert are obtained.Then we introduce the multiple correlation subjective weighting method to synthesize the evaluation results of multiple with experts’ comments,which is a more stable weighting value.So,we construct the evaluation model of combat riot control vehicle successfully.Finally,we develop the system with Python.We select four samples to illustrate the acceptability of the system by input the encode value of the morphological characteristic parameters of the samples.Compared with the results in traditional method,evaluated by experts directly,we demonstrate the system is reliable.Consequently,we provide the reference for the design of anti riot vehicles.