High-efficiency and precise identification of young apple fruits is a crucial prerequisite for automated mechanical thinning. Addressing recognition challenges caused by insufficient light, leaf obstruction, and fruit clustering, this paper proposes a recognition algorithm based on YOLO11n enhanced through dual-dimensional improvements: transfer learning and attention mechanisms. The goal is to enhance the accuracy and generalization of automated thinning machinery in complex field environments. The algorithm construction approach is as follows: build and annotate an image dataset of young apple fruits covering multiple time periods, angles, and diverse field environments to provide high-quality samples for model training; utilize transfer learning to fine-tune and optimize the YOLO11n model, enhancing its generalization capability for real-world scene recognition; incorporate the convolutional attention module into the optimally tuned baseline YOLO11n model to construct the YOLO11n-CBAM enhanced model, thereby strengthening the model's key feature extraction capabilities and target region focus abilities. Experiments demonstrate that the YOLO11n model under transfer learning achieves a precision (P) of 97.33%, a recall (R) of 88.01%, a mean average precision at 50% (mAP50) of 95.00%, and a mean average precision at 50%-95% (mAP50-95) of 69.12%. These results significantly outperform other models under transfer learning and comparable models with random weights. CBAM model further enhances YOLO11n's metrics, achieving 99.07% of precision, 88.81% of recall, 95.87% of mAP50, and 71.02% of mAP50-95, while demonstrating superior interference resistance in complex field environments. The improved model significantly enhances the accuracy of young apple fruit recognition while maintaining high generalization capabilities, providing effective technical support for automated mechanical thinning.
LIUL M, NIEL, ZHAOH L,et al.Effects of different chemical agents on flower and fruit thinning of ‘Xiahong’ ‘Huashuo’ and ‘Fuji’ apple varieties[J].Northern Horticulture,2022(21):44-49.
LIW S, WANGA L, WUZ Z,et al.Effects of fruit thinning with different chemicalson thinning and quality in Fuji apple[J].Acta Agriculturae Boreali-occidentalis Sinica,2022,31(4):468-478.
FENGB B, MEIC, WANGL,et al.Study on the effect of chemical flower and fruit thinning on Aksu Fuji apple Yanfu 3[J].Xinjiang Agricultural Sciences,2023,60(10):2470-2478.
[11]
尚忠,于建波,姜秀美,等.苹果疏果器研制[J].农业机械,2021(7):109-110.
[12]
SHANGZ, YUJ B, JIANGX M,et al.Development of an apple thinning device[J].Farm Machinery,2021(7):109-110.
ZHENGY J, ZHAIS Y, WANGY K,et al.Research advance and development trend of intelligent decision-making technology for agricultural machinery in farmland and orchard scenarios[J].Transactions of the Chinese Society of Agricultural Engineering,2025,41(18):1-27,331.
WUQ G, ZHANGW G, LIC L,et al.Segmentation of mature apple image based on Lab space under natural environment[J].Jiangsu Agricultural Sciences,2017,45(9):177-179.
HUC L, ZHAOD A, ZHAOY Y,et al.Research on localization method for overlapped apples based on maxima[J].Journal of Agricultural Mechanization Research,2016,38(3):42-46.
[23]
HASSANE, GHAZALAHA S, EI-RASHIDYN,et al.DenseNet model with attention mechanisms for robust date fruit image classification[J].International Journal of Computational Intelligence Systems,2025,18(1):228.
[24]
CONGG X, CHENX H, BINGZ Y,et al.YOLOv8-Scm:An improved model for citrus fruit sunburn identification and classification in complex natural scenes[J].Frontiers in Plant Science,2025,16:1591989.
MAB J, QIUY Y, CHENB B,et al.Apple recognition method based on improved YOLOv8n lightweight model[J].Journal of Shihezi University(Natural Science),2025,43(2):143-151.
WANGD D, HED J.Recognition of apple targets before fruits thinning by robot based on R-FCN deep convolution neural network[J].Transactions of the Chinese Society of Agricultural Engineering,2019,35(3):156-163.
SONGH B, MAB L, SHANGY Y,et al.Detection of young apple fruits based on YOLO v7-ECA model[J].Transactions of the Chinese Society for Agricultural Machinery,2023,54(6):233-242.
[31]
JIANGL L, WANGY F, WUC,et al.Fruit distribution density estimation in YOLO-Detected strawberry images:A kernel density and nearest neighbor analysis approach[J].Agriculture,2024,14(10):1848.
[32]
HANW, JIANGF, ZHUZ.Detection of cherry quality using YOLOV5 model based on flood filling algorithm[J].Foods,2022,11(8):1127.
[33]
XUW S, WANGR J.ALAD-YOLO:An lightweight and accurate detector for apple leaf diseases[J].Frontiers in Plant Science,2023,14:1204569.
LIUQ, OUYANGJ W, ZHOUZ B,et al.Segmentation and localization method for eggplants and stems based on YOLO-CRC[J].Transactions of the Chinese Society of Agricultural Engineering,2025,41(19):196-205.
SUN, SONGZ D, CAOQ,et al.Tomato leaf disease recognition method based on improved self⁃distillation EMA—DeiT and transfer learning[J].Journal of Chinese Agricultural Mechanization,2025,46(10):192-202,209.
CHENH, TANGL Q.An image classification method for apple leaf pests and diseases based on transfer learning and the VGG19 convolutional neural network[J].South Agricultural Machinery,2024,55(20):6-9.
TONGZ M, CHENX H, MAZ Y,et al.A method for detecting apple at night based on YOLOv8n with fusion of image enhancement and transfer learning[J].Journal of Huazhong Agricultural University,2024,43(5):1-9.
GUOH P, CAOY Z, WANGC S,et al.Recognition and application of apple defoliation disease based on transfer learning[J].Transactions of the Chinese Society of Agricultural Engineering,2024,40(3):184-192.
ZHOUH P, JINS X, ZHOUL,et al.Classification and recognition of Camellia oleifera fruit in the field based on transfer learning and YOLOv8n[J].Transactions of the Chinese Society of Agricultural Engineering,2023,39(20):159-166.
WUS C, MAOY M, HUH Z,et al.Detecting grape leaf diseases using improved YOLO11n[J].Transactions of the Chinese Society of Agricultural Engineering,2025,41(14):140-147.
HUANGM J, CAIW Q, ZHANGZ J,et al.Real-time accurate recognition algorithm for litchi fruit varieties based on improved YOLO11[J].Transactions of the Chinese Society of Agricultural Engineering,2025,41(11):156-164.
[50]
SAMBANAB, NNADIS H, WAJIDA M,et al.An efficient plant disease detection using transfer learning approach[J].Scientific Reports,2025,15(1):19082.
[51]
TERLAPUP V, CHANDRAN S, BANDARUR,et al.Intelligent detection of tomato crop diseases using hybrid transfer learning with PCA-Enhanced SVM classifier[J].IAENG International Journal of Computer Science,2025,52(9):2976-2991.
ZOUX B, GAOW J, SHIJ Y,et al.Food image classification technology based on improved DenseNet and transfer learning[J].Journal of Chinese Agricultural Mechanization,2025,46(6):77-84,105.
ZHANGG Z, LYUZ W, LIUH P,et al.Model for identifying lotus leaf pests and diseases using improved DenseNet and transfer learning[J].Transactions of the Chinese Society of Agricultural Engineering,2023,39(8):188-196.
HEH X, GAOX, RAOY,et al.Detection method of leaf tip in wheat seedling stage based on improved YOLOv8s[J].Scientia Agricultura Sinica,2025,58(18):3598-3615.
DONGP, WEIM H, SHIL,et al.Research and application of transfer learning in identification of maize leaf diseases[J].Journal of Chinese Agricultural Mechanization,2022,43(3):146-152.