カーブミラーを利用した交差車両の距離認知

抄録

Objects viewed through convex mirrors seem smaller than when viewed directly.　This difference causes drivers to estimate distances to crossing objects incorrectly and increases the possibility of accidents.　However, no study has determined the extent of differences in distances estimated with convex mirrors.　This study seeks to determine the extent of such differences.　We photographed 20 crossing vehicle scenes at a non-signalized intersection, 10 taken directly （direct view） and 10 taken through a convex mirror （mirror view）.　The distances from the intersection to the crossing vehicle were varied from 10 m to 32.5 m in 2.5 m intervals.　One hundred and ninety pairs of photos were generated from these twenty photos.　Pairs consisted of both direct-direct or mirror-mirror scenes and mirror-direct scenes.　Eighteen participants were asked to choose the closer scene from these pairs of photos.　If participants selected two photos at the same frequency from 190 pairs, they interpreted these photos as being almost the same distance from the vehicle.　The frequencies of closer estimation of direct and mirror views were compared.　Vehicle distances estimated from mirror views were on average 1.2 times greater than those estimated from direct views, with the greatest difference estimated from mirror views to be 1.7 times greater on average （maximum of 2.2 times greater） than those estimated from direct views.