Sage Journals: Discover world-class research

Abstract

Aiming at the key technical challenges of complex background noise interference, fruit mutual occlusion, and multi-scale object recognition in natural scene Korla pear detection tasks, an improved YoloV11 object detection algorithm integrating the Efficient Channel Attention (ECA) mechanism and Bidirectional Feature Pyramid Network (BiFPN) is proposed (ECABiFPN-YOLOv11). By introducing the ECA module to adaptively optimize feature channel weights and combining the BiFPN architecture to achieve efficient cross-level feature fusion, the model’s perception and expression capabilities for multi-scale features of Korla pear objects are significantly enhanced. The experimental results show that the improved model reaches 86.8% on the mean average precision (mAP50) index, which is 4.7 percentage points higher than that of the original YoloV11 (82.1%). The mAP@0.5:0.95 value is 62.7%, which is 4.4% higher than that of the original model. The training box_loss (final) value is 3.7% lower than that of the original model, and the verification box_loss (final) value is 3.6% lower than that of the original model. These results provide reliable technical support for the research and development of automatic grading and sorting of fragrant pear fruits and intelligent picking systems in the field of smart agriculture.

Keywords

ECA BiFPN Korla pear multi-scale features object detection

Get full access to this article

View all access options for this article.

References

Chen

Zhang

, et al. Hybrid Transformer-CNN architecture for fruit detection in overlapping scenarios. IEEE Trans Ind Electron; 69(8): 8234–8243.

Zhang

Wang

Liu

, et al. Dynamic feature fusion module for robust object detection under illumination variations. Pattern Recogn; 137: 109387.

Chen

Zhao

, et al. Neural architecture search for lightweight object detection in agriculture. J Field Robot; 41(3): 543–561.

Wang

Zhang

, et al. Cascaded attention mechanism for small object detection. IEEE Access; 9: 154321–154332.

Liu

Chen

, et al. Generative adversarial networks for low-light image enhancement in fruit detection. Comput Vis Image Understand; 228: 103687.

Zhou

Yang

, et al. Multi-modal information fusion for fruit detection in complex backgrounds. Sensors; 22(18): 6978.

Liu

Wang

, et al. Knowledge distillation for compact and efficient fruit detection models. Appl Soft Comput; 126: 109003.

Sun

Zhang

, et al. Hierarchical feature pyramid structure for multi-scale object detection. IEEE Trans Image Process; 29: 8975–8987.

Huang

Wang

Wei

, et al. Adaptive instance normalization for generalizable object detection. In: European conference on computer vision. ECCV, pp. 367–382.

10.

Guo

Zhang

Wang

, et al. Cross-layer connection module for effective feature propagation. In: IEEE international conference on image processing (ICIP). IEEE, pp. 3134–3138.

11.

Zhu

Liu

, et al. BiFPN variants for improved feature fusion in object Detection.ACM transactions on multimedia computing. Com and Appl; 17(3s): 1–19.

12.

Liu

Wang

Zhang

, et al. ECA-net enhancements for channel attention in computer vision. Neural Comput Appl; 34(15): 11849–11862.

13.

Zhang

, et al. Analysis of YoloV11’s limitations in feature fusion. J Vis Commun Image Represent; 91: 103582.

14.

Huang

Liu

Goda

. Applicability of smooth particle hydrodynamics method to large sliding deformation of saturated slopes under earthquake action. Chin J Geotech Eng 2023; 45(2): 336–344.

15.

Wang

Zhang

, et al. High-accuracy but heavyweight fruit detection models: a review. Inf Sci; 601: 287–306.

Research on a natural scene Korla pear detection method based on ECA and BiFPN improved YoloV11

Abstract

Keywords

Get full access to this article

References