基于改进Y0L0v11的水果成熟度检测

打开文本图片集

中图分类号：TP3191.4；TP183 文献标识码：A文章编号：2096-4706（2025）08-0034-07

Abstract：Aimingat theexisting problems of insufficient accuracy，the difficultyof identificationundercomplex backgrounds，ndthebvious limitationsof traditionalmethods infeatureextractioninfruitripenessdetection，afruitripenes detection algorithm （AGLU-YOLOv11） basedon improved YOLOv1 is proposed，to meet the demands for effcient data and reliable colection in fruit ripeness detection.AGLU-YOLOv11 designs the C3k2_AddBlock_CGLU module byoptimizing the C3k2 module in the YOLOv1l backbone network and integrating CATM（Conv Additive Self-Attention） and CGLU （Convolutional Gated Liear Unit），and significantlyenhances feature extraction capabilityand adaptabilityof multi-variety andmulti-stageripenessfruits.Atthesame time，theAFCAAtention Mechanism is introduced inthefeature fusion stage to strengthen global feature expresionand adaptability tocomplex backgrounds，andachieve eficient fruit quality detection andlabeling.Experimental results showthat AGLU-YOLOv1lperforsbeterin precision，robustnessand multi-saleobject adaptability than other detection models inPrecision，Recall， mAP @ 0 . 5 and 1 n A P@ 0 . 5：0 . 9 5 indicators，and can better meet the demands for identifying fruit ripeness.

Keywords：YOLO;ObjectDetection;CGLU;CATM; fruit ripenessdetection

0 引言

社会经济的飞速发展下，人们的生活水平显著提高，水果作为人类膳食结构中的重要组成部分，不仅为人体提供丰富的维生素、矿物质和膳食纤维，还在促进免疫力和预防慢性疾病中发挥着重要作用[]。（剩余9445字）