Detection of Small Apple Targets Based on Improved YOLOv5 in Natural Environments

doi:10.16182/j.issn1004731x.joss.24-0260

Abstract

Abstract:

The distribution of apples usually features occlusion and small and dense targets. To address these issues, a target detection algorithm was proposed based on an improved YOLOv5 model. Specifically, this paper added the coordinate attention (CA) mechanism, receptive field block (RFB), and adaptively spatial feature fusion (ASFF) modules to the YOLOv5, enhancing the ability to detect small targets. Additionally, the proposed algorithm replaced the CIoU in YOLOv5 with SIoU to improve the target detection box's prediction accuracy. Finally, some normal convolutions were replaced with depthwise separable convolutions (DSC), effectively reducing the calculation burden. Experiment results show that the comprehensive performance of the improved YOLOv5 is better than the original YOLOv5 and other algorithms. Moreover, its mAP value is improved by 9.6% in comparison with the original YOLOv5.

Key words: intelligent agriculture, coordinate attention mechanism, receptive field block, adaptively spatial feature fusion, small target detection, YOLOv5

CLC Number:

TP391

Liu Zilong, Zhang Lei. Detection of Small Apple Targets Based on Improved YOLOv5 in Natural Environments[J]. Journal of System Simulation, 2025, 37(8): 2124-2138.

Figures/Tables 25

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Fig. 7

Fig. 8

Fig. 9

Fig. 10

Fig. 11

Table 1

Fig. 12

Fig. 13

Fig. 14

Fig. 15

Fig. 16

Table 2

Fig. 17

Table 3

Fig. 18

Table 4

Table 5

Fig. 19

References 33

[1]	高芳芳, 武振超, 索睿, 等. 基于深度学习与目标跟踪的苹果检测与视频计数方法[J]. 农业工程学报, 2021, 37(21): 217-224.
	Gao Fangfang, Wu Zhenchao, Suo Rui, et al. Apple Detection and Counting Using Real-time Video Based on Deep Learning and Object Tracking[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(21): 217-224.
[2]	龙燕, 杨智优, 何梦菲. 基于改进YOLOv7的疏果期苹果目标检测方法[J]. 农业工程学报, 2023, 39(14): 191-199.
	Long Yan, Yang Zhiyou, He Mengfei. Recognizing Apple Targets Before Thinning Using Improved YOLOv7[J]. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(14): 191-199.
[3]	Wang Jing, Wang Leqi, Han Yanling, et al. On Combining DeepSnake and Global Saliency for Detection of Orchard Apples[J]. Applied Sciences, 2021, 11(14): 6269.
[4]	潘昕晖, 邵清, 卢军国. 基于CBD-YOLOv3的小目标检测算法[J]. 小型微型计算机系统, 2022, 43(10): 2143-2149.
	Pan Xinhui, Shao Qing, Lu Junguo. Small Object Detection Algorithm Based on CBD-YOLOv3[J]. Journal of Chinese Computer Systems, 2022, 43(10): 2143-2149.
[5]	Gao Ruilong, Zhou Qiaojun, Cao Songxiao, et al. An Algorithm for Calculating Apple Picking Direction Based on 3D Vision[J]. Agriculture, 2022, 12(8): 1170.
[6]	Zhang Chenxi, Kang Feng, Wang Yaxiong. An Improved Apple Object Detection Method Based on Lightweight YOLOv4 in Complex Backgrounds[J]. Remote Sensing, 2022, 14(17): 4150.
[7]	刘书刚, 张林坤, 杜昊东, 等. 雾天条件下改进YOLOv4的目标检测[J]. 系统仿真学报, 2023, 35(8): 1681-1691.
	Liu Shugang, Zhang Linkun, Du Haodong, et al. Improved Object Detection of YOLOv4 in Foggy Conditions[J]. Journal of System Simulation, 2023, 35(8): 1681-1691.
[8]	向南, 王璐, 贾崇柳, 等. 改进YOLO的遮挡行人检测仿真[J]. 系统仿真学报, 2023, 35(2): 286-299.
	Xiang Nan, Wang Lu, Jia Chongliu, et al. Simulation of Occluded Pedestrian Detection Based on Improved YOLO[J]. Journal of System Simulation, 2023, 35(2): 286-299.
[9]	杨文姬, 李浩, 王映龙, 等. 改进YOLOv3的多尺度高分辨率特征增强图像目标检测[J]. 小型微型计算机系统, 2023, 44(6): 1311-1317.
	Yang Wenji, Li Hao, Wang Yinglong, et al. Multi-scale High-resolution Feature Enhancement Image Target Detection Based on Improved YOLOv3[J]. Journal of Chinese Computer Systems, 2023, 44(6): 1311-1317.
[10]	Wang Dandan, He Dongjian. Channel Pruned YOLO V5S-based Deep Learning Approach for Rapid and Accurate Apple Fruitlet Detection Before Fruit Thinning[J]. Biosystems Engineering, 2021, 210: 271-281.
[11]	张稀柳, 张晓玲, 何敏军. 基于改进YOLOX-s的车辆检测方法研究[J]. 系统仿真学报, 2024, 36(2): 487-496.
	Zhang Xiliu, Zhang Xiaoling, He Minjun. Research on Vehicle Detection Method Based on Improved YOLOX-s[J]. Journal of System Simulation, 2024, 36(2): 487-496.
[12]	Liu Jiayi, Zhu Xingfei, Zhou Xingyu, et al. Defect Detection for Metal Base of TO-can Packaged Laser Diode Based on Improved YOLO Algorithm[J]. Electronics, 2022, 11(10): 1561.
[13]	Gui Zhiyong, Chen Jianneng, Li Yang, et al. A Lightweight Tea Bud Detection Model Based on Yolov5[J]. Computers and Electronics in Agriculture, 2023, 205: 107636.
[14]	Li Kangshun, Wang Jiancong, Jalil H, et al. A Fast and Lightweight Detection Algorithm for Passion Fruit Pests Based on Improved YOLOv5[J]. Computers and Electronics in Agriculture, 2023, 204: 107534.
[15]	Arifando Rio, Eto Shinji, Wada Chikamune. Improved YOLOv5-based Lightweight Object Detection Algorithm for People with Visual Impairment to Detect Buses[J]. Applied Sciences, 2023, 13(9): 5802.
[16]	Guo Jinkai, Xiao Xiao, Miao Jianchi, et al. Design and Experiment of a Visual Detection System for Zanthoxylum-harvesting Robot Based on Improved YOLOv5 Model[J]. Agriculture, 2023, 13(4): 821.
[17]	Sun Han, Wang Bingqing, Xue Jinlin. YOLO-P: An Efficient Method for Pear Fast Detection in Complex Orchard Picking Environment[J]. Frontiers in Plant Science, 2023, 13: 1089454.
[18]	Yao Jia, Qi Jiaming, Zhang Jie, et al. A Real-time Detection Algorithm for Kiwifruit Defects Based on YOLOv5[J]. Electronics, 2021, 10(14): 1711.
[19]	Hu Jinnan, Li Guo, Mo Haolan, et al. Crop Node Detection and Internode Length Estimation Using an Improved YOLOv5 Model[J]. Agriculture, 2023, 13(2): 473.
[20]	Li Shilin, Zhang Shujuan, Xue Jianxin, et al. Lightweight Target Detection for the Field Flat Jujube Based on Improved YOLOv5[J]. Computers and Electronics in Agriculture, 2022, 202: 107391.
[21]	He Jiaxing, Wang Xiaodan, Song Yafei, et al. A Multiscale Intrusion Detection System Based on Pyramid Depthwise Separable Convolution Neural Network[J]. Neurocomputing, 2023, 530: 48-59.
[22]	Huo Yinuo, Zhang Qixing, Jia Yang, et al. A Deep Separable Convolutional Neural Network for Multiscale Image-based Smoke Detection[J]. Fire Technology, 2022, 58(3): 1445-1468.
[23]	Ma Jie, Lu Ange, Chen Chen, et al. YOLOv5-lotus an Efficient Object Detection Method for Lotus Seedpod in a Natural Environment[J]. Computers and Electronics in Agriculture, 2023, 206: 107635.
[24]	Zhu Ruilin, Zou Hongyan, Li Zhenye, et al. Apple-net: A Model Based on Improved YOLOv5 to Detect the Apple Leaf Diseases[J]. Plants, 2023, 12(1): 169.
[25]	Liu Zhiwei, Gan Menghan, Xiong Li, et al. Multilevel Receptive Field Expansion Network for Small Object Detection[J]. IET Image Processing, 2023, 17(8): 2385-2398.
[26]	Hrishikesh P S, Puthussery D, Akhil K A, et al. Relativistic GAN Using Receptive Field Block for Single Image Super-resolution with Improved Perceptual Quality[C]//2023 11th International Symposium on Electronic Systems Devices and Computing (ESDC). Piscataway: IEEE, 2023: 1-6.
[27]	Qian Yukun, Miao Yalun, Huang Shuqin, et al. Real-time Detection of Eichhornia Crassipes Based on Efficient YOLOV5[J]. Machines, 2022, 10(9): 754.
[28]	Xue Zhenyang, Xu Renjie, Bai Di, et al. YOLO-tea: A Tea Disease Detection Model Improved by YOLOv5[J]. Forests, 2023, 14(2): 415.
[29]	Li Yaodi, Xue Jianxin, Zhang Mingyue, et al. YOLOv5-ASFF: A Multistage Strawberry Detection Algorithm Based on Improved YOLOv5[J]. Agronomy, 2023, 13(7): 1901.
[30]	Ji Wei, Pan Yu, Xu Bo, et al. A Real-time Apple Targets Detection Method for Picking Robot Based on ShufflenetV2-YOLOX[J]. Agriculture, 2022, 12(6): 856.
[31]	Zhang Wei, Xia Xulu, Du Jianming, et al. Recognition and Detection of Wolfberry in the Natural Background Based on Improved YOLOv5 Network[C]//2022 3rd International Conference on Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA). Piscataway: IEEE, 2022: 256-262.
[32]	Zhu Shisong, Ma Wanli, Wang Jianlong, et al. EADD-YOLO: An Efficient and Accurate Disease Detector for Apple Leaf Using Improved Lightweight YOLOv5[J]. Frontiers in Plant Science, 2023, 14: 1120724.
[33]	Gevorgyan Zhora. SIoU Loss: More Powerful Learning for Bounding Box Regression[EB/OL]. (2022-05-25) [2024-03-10]. .

DSC数量	参数量	P/%	R/%	mAP0.5/%
8	13 395 370	91.4	82.9	90.3
9	13 387 281	92.0	83.4	90.6
10	13 256 465	91.7	82.8	90.2
11	12 732 689	92.3	83.8	91.0
12	12 471 057	87.8	77.0	85.6
13	11 423 505	86.7	75.7	84.4
14	11 260 433	86.1	75.0	83.5

模型	P	R	mAP0.5
YOLOv5	84.6	72.2	81.4
改进YOLOv5	92.3	83.8	91.0

CA	RFB	ASFF	SIoU	P	R	mAP0.5
				84.6	72.2	81.4
√				83.8	72.7	81.4
	√			89.4	79.6	87.1
		√		89.0	78.8	86.9
			√	84.2	72.8	81.4
√	√	√	√	92.3	83.8	91.0

模型	P	R	mAP0.5
Faster RCNN	57.83	54.75	48.22
SSD	77.78	8.82	31.70
Efficientdet	97.77	11.45	15.92
YOLOv4	81.60	67.75	74.96
YOLOv5	84.60	72.20	81.40
改进YOLOv5	92.30	83.80	91.00