Visual Perception And Detection Of A Small Unmanned Aerial Vehicle

For safety and security reasons,DJI-Innovations released the polygon no-fly strategy,which took effect at China’s airports first.Unmanned aerial vehicle(UAV)invasions into prohibited areas are becoming more and more common,despite no-fly zone(NFZ)and restricted fly zone(RFZ)have been announced.The literature has shown little information concerning the capability of humans on the ground in apperceiving and detecting the invasion of a UAV into the airspace,although this capability was important.In view of this,this study determines the influencing factors and variation rules of observer perception and detection by means of two field experiments on UAVs.In the part of visual perception behavior research,the means of threshold measurement was used to study the influence factors and change rules of small UAV perception capability of ground observers.Twenty-four participants were asked to rate their perception of a drone hovering in airspace on a five-point scale and to give a fourpoint auditory scale.The linear distance threshold and visual angle threshold were calculated using the subjective visual perception score.A total of 384 groups of data were collected and processed by ANOVA,t-test and Logistic regression.The results showed that the subjects could hardly hear the sound when the drone was hovering at the assigned location.The auditory rating of 1(it means the voice was heard completely clearly)or 2(it means the voice was heard clearly)did not appear.Both view angle and linear distance had significant influence on the probability of scoring.The probability showed a positive correlation with the angle and a negative correlation with the distance.The effect of the frame on the probability of scoring was not significant.The visual line-of-sight distance had defined by The Civil Aviation Administration of China was500 m,which was not universal.In the part of visual detection behavior research,the signal detection theory(SDT)was used to study the influencing factors of small UAV invasion detection capability of ground observers.Twenty-four participants were asked to judge whether the drone hovering in the air near the borderline had invaded the test site on a five-point scale and to give a four-point auditory scale.The discrimination(d’),the area of receiver’s operating characteristics(ROC)curves(P(A))and the response bias(bias)were calculated using the subjective visual detection score.A total of 1728 groups of scoring tasks were carried out in the detection experiment,and the data obtained were processed by ANOVA and correlation analysis.The results show a positive correlation between auditory score and detection score.The distance between the subjects and the test site boundary had significant influence on the P(A),bias and d’.It showed a negative correlation with the P(A)and d’ and a positive correlation with the bias.The protection frame had a significant impact on bias,and the bias was smaller when the drone is equipped with a protection frame.The horizontal distance of the UAV to the boundary had a significant influence on P(A)and d’.It was positively correlated with both of them.The groups and observation time of the subjects had no significant effect on the three indexes.The research results of this paper can be used to master the understanding degree of UAV position in the process of visual perception and detection,reduce the occurrence of UAV accidents.It can also be used to train relevant staff and improve work efficiency.Meanwhile,the experimental conclusion can provide reference for UAV related regulation.The thesis includes 24 Figure s,30 tables and 97 references in this paper.