Article Details

Abstract

To address the challenges in apple fruit detection under natural orchard environments, including large variations in fruit scale, complex illumination conditions, and the tendency for small targets to be missed, an apple fruit detection algorithm based on YOLO11n (YOLO11n-AL) is proposed. In the backbone network, the original C3k2 module is improved by introducing a C3k2_AP module integrated with the APBottleneck attention mechanism. By leveraging asymmetric padding and multi-directional feature modeling, the feature representation capability is enhanced, thereby improving the extraction of apple fruit features. During the feature fusion and detection stages, an improved Detect_LQE module is introduced, which explicitly models classification confidence through a label quality evaluation mechanism, alleviating the inconsistency between localization accuracy and classification scores and effectively enhancing detection stability. Experimental results show that, compared with the original YOLO11n model, the proposed method achieves improvements of 1.4%, 2.3%, 3.8%, and 3.2% in precision, recall, mAP@0.5, and mAP@0.5:0.95, respectively, while maintaining a relatively low parameter scale, demonstrating its suitability for accurate apple fruit detection in orchard environments.

Keywords

YOLO11n; apple fruit detection; object detection; attention mechanism; deep learning

References

1. [1] Huang Zhihao, Lu Chengfang, Cui Yanrong, etc YOLO-AP: A Lightweight Apple Fruit Detection Algorithm Based on Improved YOLO11n [J]. China Agricultural Science and Technology Review, 2025, 27 (10): 118-133
2. [2] Jakhfer A，Tsvetelina G，Eleonora N，et al. Deep Learning-Based Identification of Kazakhstan Apple Varieties Using Pre-Trained CNN Models [J/OL].AgriEngineering，2025，7（10）：331（2025-10-01）.https://doi.org/10.3390/agriengineering7100331.
3. [3] Qian J，Wei L J. Apple leaf disease image recognition based on a modified rime optimization algorithm and ConvNeXt network [J/OL].Frontiers in Plant Science，2025，16：1626335（2025-09-10）.https://doi.org/10.3389/fpls.2025.1626335.
4. [4] Zhu Licheng, Wang Wenbei, Zhao Bo, etc Detection method of low rootstock dense planting orchard apples based on SDE-YOLO [J]. Journal of Agricultural Machinery, 2025, 56 (9): 638-647
5. [5] Wu Zhenchao, Yang Yu, Shi Mengqing, etc Research progress on thinning techniques during the transition from traditional apple orchards to modern apple orchards [J]. Modern Agricultural Equipment, 2026, 47 (1): 1-11
6. [6] Bhattacharya S，Arunachalam C，Zhang K X，et al. Detect, segment, cluster: Apple localization for robotic harvesting in complex orchards [J/OL].Smart Agricultural Technology，2025，12：101642.https://doi.org/10.1016/j.atech.2025.101642.
7. [7] Li specializes in research Research on Apple Recognition and Localization Technology in Orchard Environment Based on Deep Learning [D]. Harbin: Harbin University of Commerce, 2025
8. [8] Xu N，Ning X S，Li A J，et al. Research on Orchard Navigation Line Recognition Method Based on U-Net [J/OL].Sensors，2025，25（22）：6828.https://doi.org/10.3390/s25226828.
9. [9] Tang Zhiyong, He Yichuan, Tang Zhihui, etc Research Status and Development Trends of Orchard Picking Machinery [J]. Zhejiang Agricultural Journal, 2026, 38 (1): 184-196
10. [10] Wen Binbin, Zhang Hua, Meng Xianglong Lightweight apple detection method based on improved YOLO v5 [J]. Jiangsu Agricultural Science, 2024, 52 (12): 217-223
11. [11] Wang C，Guo X R，Ma C Y，et al. A Flat Peach Bagged Fruits Recognition Approach Based on an Improved YOLOv8n Convolutional Neural Network [J/OL].Horticulturae，2025，11（11）：1394.https://doi.org/10.3390/horticulturae11111394.
12. [12] Gao Yanqing, Xu Xiaoma, Liu Guangchao, etc Foreign object detection method for conveyor belt based on improved YOLOv11n model [J]. Industrial and Mining Automation, 2025, 51 (10): 41-47, 56
13. [13] Xu Yongqi Research on Textile Material Classification and Defect Recognition Technology Based on Deep Learning [D]. Nanjing: Nanjing University of Information Science and Technology, 2025
14. [14] Wei Ying'ao, Xu Jian, Li Yinghao, etc Abnormal detection of substation equipment and production behavior based on improved YOLOv11n model [J]. Journal of Hubei University for Nationalities (Natural Science Edition), 2025, 43 (1): 41-46
15. [15] Jin Huawei, Weng Changkun Foreign Object Detection Algorithm for Coal Mine Conveyor Belt Based on LDAFL-YOLOv11 [J/OL]. Coal Science and Technology, 2025:1-4 https://link.cnki.net/urlid/11.2402.td.20251201.1504.004.

Funding

Henan Province Science and Technology Research Project (252102221017); Natural Science Foundation of Zhengzhou University of Science and Technology (2025-ZKYZQN-02)

About the authors

Xiong Du (1998-), male, Han ethnicity, from Xinyang, Henan, intermediate engineer, master's degree, research direction: machine vision.