基于轻量化网络的无人机图像目标识别

打开文本图片集

引用格式：，，，等.基于轻量化网络的无人机图像目标识别[J].指挥控制与仿真，2025，47（6）：76-81.GUOJ，GUOJ，MAY，etalsdolgteightokoeiageaetcoiionJoadorol&iuatio5，4（6）61.

Abstract：Inresponse totheproblemoflargecomputationalcomplexityandfalsedetectionsintargetdetectionbydrones.A targetrecognition method basedonlightweight networks has been proposed.Basedonthe YOLOv5object detectionalgorithm，thealgorithmhasbeenoptimizedusing theFasterNetlightweightnetworkarchitecture，whichreducesthenumberof network parametersandimproves theeficiencyof thealgorithm.Inordertoaccuratelycaptureandemphasizekeyinformationinteinput sequence，andfurther enhancetheperformanceofthealgorithm，aparameterfreeattention mechanism SimAMis introduced.Theresultsindicatethatthismethod isanoptimizedapplicationofobjectdetectiontechnology，which canbeterbalance therelationshipbetweendetectionspeedandaccuracy，andachieve beterdetectionresults inunmanned aerial vehicle aerial image detection tasks.

KeyWords：object detection；YOLOv5；lightweight network structure；attention mechanism

无人机技术在近年来取得了显著的进步与发展，这一趋势在俄乌冲突与巴以冲突中得到了鲜明的体现。（剩余9296字）