基于NanoDet-f的轻量级公交站点客流检测网络

打开文本图片集

中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2025）18-0035-05

Abstract： Object detectionalgorithms are widelyused in fieldsof surveillnce，autonomous driving，and industrial robotics，butteomputatioalndorostratsofdgeicesecallengestodelgtweightuacydd Tothis ed，thisstudyproposes thelightweightobject detection modelNanoDet-f，whichemploysanAtention Mechanismand aninovativeFANfeaturefusionmodule toimproveacuracywhileensuring effcientiference.Theexperimentsareconducted onNVIDIARTX3090，and theresultsshow thattheAP @0.5 of the model reaches 85.9% ，andthe inferencespeedisfast，which is suitable forthedeploymentofedgedevices.Ithasanimportantreference valueforthedeploymentof lightweight models in real engineering applications.

Keywords：Object Detection; Attention Mechanism; lightweight model; NanoDet-f

0 引言

在当今社会，公共交通系统的高效运行对城市发展至关重要。（剩余7034字）