Low-altitude Target Detection And Identification And Threat Assessment In Haze Weather

In the wake of 911 in the United States,security threats from the air have attracted increasing attention.The common low-altitude targets such as drones and hot air balloons are small in size,low in cost,simple in operation,and numerous in variety,and pose a great threat to public safety.In recent years,with the frequent occurrence of haze weather,it is more urgent to prevent and control low-altitude targets.Therefore,this paper proposes a method of lowaltitude target recognition and threat assessment under haze weather.Firstly,the characteristics of low altitude targets were studied,and the frequent haze weather and complex background on the ground were analyzed.According to the low visibility of haze weather,various haze removal algorithms were studied.Based on this,a haze removal method based on single gray image restoration was proposed,and compared with the existing haze removal methods.It was proved that this method had a good effect on the quality of haze removal and scene fidelity.Secondly,we designed a low-altitude target detection and recognition scheme combining data driven and knowledge driven to realize low-altitude target location and recognition.The basic shape characteristics,affine invariant moment characteristics and HOG features were integrated into the feature vectors as parameters to represent the target characteristics,so as to achieve the effect of complementary advantages between different features.The separation of sky and ground was realized by using horizontal grayscale projection,and the detection and positioning of suspicious targets were realized by combining horizontal mixed grayscale gradient index and vertical mixed grayscale gradient index.In addition,the variation coefficient method was introduced into the target recognition field to achieve feature reduction.In the end,the method of tracking and threat assessment for identifying the located lowaltitude targets was studied,and the concentration of haze,the type,distance and speed of lowaltitude targets,and whether low-altitude targets could be answered as indicators of the threat level of the target were used.Using expert evaluation and fuzzy logic to obtain the index value,the distance and speed calculation method of low altitude target was calculated in detail.The ideal method of multi-attribute decision-making was adopted to evaluate the threat to key areas and important places for low-altitude targets,and to maintain public safety by automatically starting early warning and asking for help.