School of Electrical & Information Engineering, Jiangsu University, Zhenjiang 212013, China
Show less
文章历史+
Received
Published
Revised
2025-02-08
2026-01-18
2025-10-14
Online Date
2026-09-20
PDF (23178K)
摘要
为实现并联机器人分拣系统在大视场下高精度识别检测堆叠果串果梗位置,提出了一种基于改进YOLOv5s模型的小目标低对比度检测方法(CNC-YOLOv5s)。针对大视场下的果梗小目标,通过将YOLOv5s模型主干网络中的C3卷积改进为具有并联残差结构的C3-MSE卷积,并通过复用方式将不同卷积层提取到的特征信息进行融合,以解决YOLOv5s模型存在提取果梗纹理特征信息不足问题;在堆叠果串中识别检测低对比度果梗位置时,YOLOv5s模型由于学习时权重分配平均,对目标关键特征关注不够,存在提取果梗边缘层次特征信息不足问题,通过在其主干网络后3个C3卷积层中嵌入坐标注意力模块(coordinate attention, CA),并在CA中增加最大池化层,以动态调整YOLOv5s模型对目标特征学习的权重,增强对目标边缘层次特征信息的关注;针对因模型结构改进引起YOLOv5s模型收敛速度变慢问题,通过将YOLOv5s模型头部网络中的交并比(Complete Intersection over Union, CIoU)回归损失函数改进为不受目标边界框偏差影响的函数(NWD),以加速YOLOv5s模型收敛。实验结果表明,改进后的CNC-YOLOv5s模型mAP值达95.64%,较原模型提升4.79%,与SSD、Faster RCNN、YOLOv4、YOLOvX等主流模型相比,mAP值分别提升19.4%、13.02%、7.33%、3.11%,验证了所提方法对堆叠果串果梗位置检测的优越性,能更好满足并联机器人分拣要求,可为葡萄等水果的产后智能化分拣提供技术参考。
Abstract
To realize the parallel robot sorting system to recognize and detect the location of fruit stalks of stacked fruit bunches with high accuracy under a large field of view, a small target low-contrast detection method based on CNC-YOLOv5s (improved YOLOv5s model) was proposed. For the small target of fruit stalks under the large field of view, the C3 convolution in the backbone network of YOLOv5s model was improved to C3-MSE convolution with parallel residual structure, and the feature information extracted from different convolution layers was fused by multiplexing to solve the problem of insufficient extracted texture feature information of fruit stalks in YOLOv5s model. When recognizing and detecting the location of low-contrast fruit stalks in stacked fruit bunches, the YOLOv5s model pays insufficient attention to the key features of the target due to the average distribution of weights during learning, and there is a problem of inadequate extraction of edge-level feature information of the stalks, which is solved by embedding CA (coordinate attention) module in the last three C3 convolutional layers of the backbone network and adding a Maximum pooling layer, to dynamically adjust the weights of YOLOv5s model for target feature learning, and enhance the attention to the target edge level feature information. To address the problem of slower convergence of the YOLOv5s model caused by the improvement of the model structure, CIoU (complete intersection over union) regression loss function in the head network of the YOLOv5s model was improved to one that is unaffected by the deviation of the target bounding box NWD to accelerate the convergence of the YOLOv5s model. The experimental results show that the mAP value of the improved CNC-YOLOv5s model reaches 95.64%, which is 4.79% higher than that of the original model. The mAP value is 19.4%, 13.02%, 7.33%, and 3.11% higher than that of the mainstream models, such as SSD, Faster RCNN, YOLOv4, and YOLOvX, etc., which verifies the effectiveness of the proposed method for the superiority of the proposed method to detect the position of fruit stalks in stacked bunches can better meet the sorting requirements of parallel robots, which can provide technical reference for the post-production intelligent sorting of grapes and other fruits.
YI Cheng-guo, GAO Guo-qin.
Position Detection Method of Cluster Fruit Stalk of Sorting Parallel Robot[J].
Science Technology and Engineering, 2026, 26(2): 680-692 DOI:10.12404/j.issn.1671-1815.2500759
ZhangQian, GaoGuoqin. Parallel robot string fruit grasping model and grasping pose calculation[J]. Transactions of the Chinese Society of Agricultural Engineering, 2019, 35(23): 37-47.
[3]
LawalM O. Tomato detection based on modified YOLOv3 framework[J]. Scientific Reports, 2021, 11(1): 1-11.
[4]
SunM, XuL, LuoR, et al. GHFormer-Net: Towards more accurate small green apple/begonia fruitdetection in the nighttime[J]. Journal of King Saud University-Computer and Information Sciences, 2022, 34(7): 4421-4432.
PengHongxing, LiJing, XuHuiming, et al. Litchi detection based on multiple feature enhancement and feature fusion SSD[J]. Tran-sactions of the Chinese Society of Agricultural Engineering, 2022, 38(4): 169-177.
[7]
MaiX, ZhangH, JiaX, et al. Faster R-CNN with classifier fusion for automatic detection of small fruits[J]. IEEE Transactions on Automation Science and Engineering, 2020, 17(3): 1555-1569.
ZhaoXiaoyan, FangJiandong, ZhaoYudong. Improved tomato seedling classification and recognition model of YOLOv5s[J]. Science Technology and Engineering, 2024, 24(27): 11774-11785.
[10]
GuoT, XuX. Salient object detection from low contrast images based on local contrast enhancing and non-local feature learning[J]. The Visual Computer, 2021, 37(8): 2069-2081.
HaoJianjun, BingZhenkai, YangShuhua, et al. Improved YOLOv3 algorithm for detection of green walnut[J]. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(14): 183-190.
[13]
LuS, ChenW, ZhangX, et al. Canopy-attention-YOLOv4-based immature/mature apple fruit detection on dense-foliage tree architectures for early crop load estimation[J]. Computers and Electro-nics in Agriculture, 2022, 193: 106-109.
HanYajie, HaoXiaoli, NiuBaoning, et al. Conveyor belt damage detection method with small size and low contrast[J]. Journal of Taiyuan University of Technology, 2022, 32(27): 1-10.
HuZirui, DingJianwei, TianBowen. Retinex low-light image enhancement algorithm based on residual and multi-scale attention mechanism[J]. Science Technology and Engineering, 2024, 24(35): 15118-15125.
[21]
ZhangQ, GaoG Q. Prioritizing robotic grasping of stacked fruit clusters based on stalk location in RGB-D images[J]. Computers and Electronics in Agriculture, 2020, 172: 1-14.
[22]
ZhangQ, GaoG Q. Hand-eye calibration and grasping pose calculation with motion error compensation and vertical-component correction for 4-R(2-SS) parallel robot[J]. International Journal of Advanced Robotic Systems, 2020, 17(2): 1-14.
[23]
JiangT, LiC, YangM, et al. An improved YOLOv5s algorithm for object detection with an attention mechanism[J]. Electronics, 2022, 11(16): 421-430.
[24]
ZengN, WuP, WangZ, et al. A small-sized object detection oriented multi-scale feature fusion approach with application to defect detection[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-14.
[25]
GaoS H, ChengM M, ZhaoK, et al. Res2net: a new multi-scale backbone architecture[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 43(2): 652-662.
[26]
LiH, LiC, LiG, et al. A real-time table grapedetection method based on improved YOLOv4-tiny network in complex background[J]. Biosystems Engineering, 2021, 212: 347-359.
[27]
OuZ, WangZ, XiaoF, et al. AD-RCNN: Adaptive dynamic neural network for small object detection[J]. IEEE Internet of Things Journal, 2022, 10(5): 4226-4238.
[28]
CaoY, WuY, LiM, et al. DFAF-Net: a dual-frequency PolSAR image classification network based on frequency-aware attention and adaptive feature fusion[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 1-18.
[29]
XuC, WangJ, YangW, et al. Detecting tiny objects in aerial images: a normalized Wasserstein distance and a new benchmark[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2022, 190: 79-93.