改进YOLOv8n的花生品质检测方法

doi:10.3778/j.issn.1002-8331.2407-0060

摘要/Abstract

摘要： 花生品质筛选在农业生产和食品安全中具有重要意义。针对传统花生品质筛选方法效率低的问题，提出改进YOLOv8n算法的轻量化花生品质检测模型LE-YOLO（lightweight and efficient）。提出一种分组重序颈部模块（grouped shuffling bottleneck，GSBottleneck），增加了模型非线性拟合能力，减少了模型推理时间；设计了残差分组重序模块（residual group shuffling block，ResGSBlock），并结合GSConv（grouped shuffle convolution）构建轻量颈部网络（lightweight neck，LW-Neck），减少了模型计算成本，提升了模型推理速度；提出自适应特征优化模块（adaptive feature optimization block，AFOB），增强了通道间信息交互和模型表征能力。在DW花生数据集上进行实验验证，相较于YOLOv8n算法，LE-YOLO的计算量减少了1?GFlops，FPS提升了25%，平均精度均值mAP@0.5达到了98%，验证了该算法在检测精度和速度上的良好性能，为花生品质筛选提供了一种有效的方法。

关键词: YOLOv8n, GSConv, GSBottleneck, 花生品质筛选, 轻量化模型

Abstract: Peanut quality screening is of great significance in agricultural production and food safety. In order to solve the problem of low efficiency of traditional peanut quality screening methods, a lightweight peanut quality detection model LE-YOLO (lightweight and efficient) with improved YOLOv8n algorithm is proposed. A grouped shuffling bottleneck (GSBottleneck) module is proposed, which increases the nonlinear fitting ability of the model and reduces the model inference time. A residual group shuffling block (ResGSBlock) is designed, and a lightweight neck (LW-Neck) is constructed by combining GSConv (grouped shuffle convolution), which reduces the cost of model calculation and improves the inference speed of the model. An adaptive feature optimization block (AFOB) is proposed to enhance the information interaction and model characterization capabilities between channels. Experimental verification on the DW peanut dataset shows that compared with the YOLOv8n algorithm, the computational cost of LE-YOLO is reduced by 1 GFlops, the FPS is increased by 25%, and the average accuracy reaches 98% mAP@0.5, which verifies the good performance of the algorithm in detection accuracy and speed, and provides an effective method for peanut quality screening.

Key words: YOLOv8n, grouped shuffle convolution (GSConv), grouped shuffling bottleneck (GSBottleneck), peanut quality screening, lightweight model

黄英来, 牛达伟, 侯畅, 杨柳松. 改进YOLOv8n的花生品质检测方法[J]. 计算机工程与应用, 2024, 60(23): 257-267.

HUANG Yinglai, NIU Dawei, HOU Chang, YANG Liusong. Improved Peanut Quality Detection Method of YOLOv8n[J]. Computer Engineering and Applications, 2024, 60(23): 257-267.

参考文献

[1] HE S, CHEN Y, XIANG W, et al. Carbon and nitrogen footprints accounting of peanut and peanut oil production in China[J]. Journal of Cleaner Production, 2021, 291: 125964.
[2] 陈立辛, 王磊, 乔印虎, 等. 基于深度学习的花生米缺陷识别分拣方法研究[J]. 包装与食品机械, 2022, 40(3): 65-70.
CHEN L X, WANG L, QIAO Y H, et al. Deep learning-based peanut rice appearance defect identification and sorting method study[J]. Packaging and Food Machinery, 2022, 40(3): 65-70.
[3] 赵志衡, 宋欢, 朱江波, 等. 基于卷积神经网络的花生籽粒完整性识别算法及应用[J]. 农业工程学报, 2018, 34(21): 195-201.
ZHAO Z H, SONG H, ZHU J B, et al. Identification algorithm and application of peanut kernel integrity based on convolution neural network[J]. Transactions of the Chinese Society of Agricultural Engineering, 2018, 34(21): 195-201.
[4] DENG L, HAN Z. Image features and DUS testing traits for peanut pod variety identification and pedigree analysis[J]. Journal of the Science of Food and Agriculture, 2019, 99(5): 2572-2578.
[5] VELESACA H O, MIRA R, SUáREZ P L, et al. Deep learning based corn kernel classification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020: 66-67.
[6] BAL F, KAYAALP F. A novel deep learning-based hybrid method for the determination of productivity of agricultural products: apple case study[J]. IEEE Access, 2023, 11: 7808-7821.
[7] GULZAR Y. Fruit image classification model based on MobileNetV2 with deep transfer learning technique[J]. Sustainability, 2023, 15(3): 1906.
[8] BUTUNER R, CINAR I, TASPINAR Y S, et al. Classification of deep image features of lentil varieties with machine learning techniques[J]. European Food Research and Technology, 2023, 249(5): 1303-1316.
[9] 王春龙, 蒋仲铭, 鲍安红. 基于协调注意力的花生荚果品质分级[J]. 食品与机械, 2022, 38(9): 180-184.
WANG C L, JIANG Z M, BAO A H. Classification of peanut quality based on coordinated attention[J]. Food and Machinery, 2022, 38(9): 180-184.
[10] YANG H, NI J, GAO J, et al. A novel method for peanut variety identification and classification by Improved VGG16[J]. Scientific Reports, 2021, 11(1): 15756.
[11] WANG Y, DING Z, SONG J, et al. Peanut defect identification based on multispectral image and deep learning[J]. Agronomy, 2023, 13(4): 1158.
[12] SUN G, WANG S, XIE J. An image object detection model based on mixed attention mechanism optimized YOLOv5[J]. Electronics, 2023, 12(7): 1515.
[13] TERVEN J, CóRDOVA-ESPARZA D M, ROMERO-GONZáLEZ J A. A comprehensive review of yolo architectures in computer vision: from yolov1 to yolov8 and yolo-nas[J]. Machine Learning and Knowledge Extraction, 2023, 5(4): 1680-1716.
[14] ZHANG Z, HE T, ZHANG H, et al. Bag of freebies for training object detection neural networks[J]. arXiv:1902. 04103, 2019.
[15] LIU H, HOU Y, ZHANG J, et al. Research on weed reverse detection methods based on improved you only look once (YOLO)v8: preliminary results[J]. Agronomy, 2024, 14(8): 1667.
[16] XIAO B, NGUYEN M, YAN W Q. Fruit ripeness identification using YOLOv8 model[J]. Multimedia Tools and Applications, 2024, 83(9): 28039-28056.
[17] LUO B, KOU Z, HAN C, et al. A “hardware-friendly” foreign object identification method for belt conveyors based on improved YOLOv8[J]. Applied Sciences, 2023, 13(20): 11464.
[18] WU Y, HAN Q, JIN Q, et al. LCA-YOLOv8-Seg: an improved lightweight YOLOv8-Seg for real-time pixel-level crack detection of dams and bridges[J]. Applied Sciences, 2023, 13(19): 10583.
[19] ZHANG H, WANG Y, DAYOUB F, et al. VarifocalNet: an IoU-aware dense object detector[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 8514-8523.
[20] 马阿辉, 祝双武, 李丑旦, 等. 改进YOLOv5的织物疵点检测算法[J]. 计算机工程与应用, 2023, 59(10): 244-252.
MA A H, ZHU S W, LI C D, et al. Improved fabric defect detection algorithm of YOLOv5[J]. Computer Engineering and Applications, 2023, 59(10): 244-252.
[21] BIE M, LIU Y, LI G, et al. Real-time vehicle detection algorithm based on a lightweight you-only-look-once (YOLOv5n-L)approach[J]. Expert Systems with Applications, 2023, 213: 119108.
[22] LI C, LI L, JIANG H, et al. YOLOv6: a single-stage object detection framework for industrial applications[J]. arXiv:2209.02976, 2022.
[23] WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 7464-7475.
[24] WU T, DONG Y. YOLO-SE: improved YOLOv8 for remote sensing object detection and recognition[J]. Applied Sciences, 2023, 13(24): 12977.
[25] WANG G, CHEN Y, AN P, et al. UAV-YOLOv8: a small-object-detection model based on improved YOLOv8 for UAV aerial photography scenarios[J]. Sensors, 2023, 23(16): 7190.