基于改进YOLOv8模型的轻量化梨果实检测

打开文本图片集

，.基于改进YOLO v8模型的轻量化梨果实检测[J].江苏农业学报，2026，42（3）：582-590.doi：10.3969/j.issn.1000-4440.2026.03.016

关键词：梨；果实；YOLO Δv8 模型；检测精度；检测速率；计算量；参数量中图分类号：TP391.4 文献标识码：A 文章编号： 1000-4440（2026）03-0582-09

Lightweight pear fruit detection based on an improved YOLO v8 model

LUO Yuntao， ZHANG Zhi'an （ 21oo94，China）

Abstract：Toaddressthe challenges poor detection performance target detection models for pear-picking robots in complexorchardenvironmentsandthediffcultydeploymentonresource-constrainedembeddedplatforms，alightweightpear fruit detection modelbasedonanimprovedYOLOv8isproposedinthisstudy.First，anovelPR_Botleneck module was introduced，whichemploedaninovativestructuraldesigntoefectivelyducethemodel’sparametercountandcomputatioalload， significantly easingtheoperational burden.Second，the SPPF_CBAMmodule was proposed to greatlyenhancethe model'sfocus ontargetregions，therebysignificantlyimprovingdetectionaccuracyincomplexscenarios.ThemodelalsoincoporatedanADown downsamplingmodule，whichnotonlyreducedmodelcomplexityandparametercountbutalsostrengthened informationretention duringthe downsampling process.Finally，the WIoUloss function wasintroduced，which enabled the model toconcentrateon medium-qualityanchorboxesduringtraining，thusimprovingoveralldetectionperformanceandenhancingthemodel’sgeeralizationcapability.ExperimentalresultsshowedthatcomparedwiththeoriginalYOLOv8model，theprecision，recall，andmean average precision （ mAP50 ） the improved model increased byO.7，3.7，and1.7 percentage points，respectively.The parameter count and computational load were reduced by 38.72% and 30.86% ，respectively. The model size was reduced by 38.33 % ，andthe detection speed reached 85.4 frames per second，an improvement 12.8 frames per second over the original model.Therefore，the improved YOLOv8 lightweight model not only improves detection accuracy and speed，but also significantly decreases the computational load and parameter count，enablingrapid and effectivereal-timepearfruitdetectiononembeddedplatforms.

Key words：pear；fruit；YOLO v8 model；detection accuracy；detection speed;computational load；parameter count

梨果实具有润肺止咳、促进消化、补充水分等功效[1]。（剩余13502字）