Design of Pineapple Recognition System Based on Deep Learning

DOI: https://doi.org/10.62517/jike.202604135

Author(s)

Wang Runtao*, Li Zelong, Huang Chuqi, Chen Xiaoran, Liang Rongwei, Lin Zixin, Zeng Fanjin, Li Zhitong, Zhang Qingyi

Affiliation(s)

School of Electronic and Electrical Engineering, Lingnan Normal University, Guangdong, China *Corresponding author

Abstract

In view of the current situation that pineapple picking operation is highly dependent on manpower and the level of automation is insufficient, this study designed a set of pineapple fruit automatic recognition system suitable for complex background based on deep learning and embedded technology. The hardware of the system is composed of a target recognition module and a control module. The target recognition module uses k210 as the main control chip to realize the real-time acquisition and processing of pineapple images. By comparing the performance of the algorithm, yolov5 performs best in the pineapple recognition task, and the mean accuracy (map) is 0.987. The benchmark yolov5 model was further improved in two aspects: first, the collaborative attention mechanism (CA) was introduced to improve the map to 0.992; The second is to use the lightweight backbone network shufflenet V2 and optimize the structure, which significantly reduces the amount of model parameters. This research provides an effective technical scheme for the automation and intelligence of pineapple production.

Keywords

Pineapple Object Detection; Deep Learning; Algorithm Optimization; K210; Attention Mechanism; Model Lightweighting

References

[1]Carolina Trentin, Yiannis Ampatzidis, et al. Tree crop yield estimation and prediction using remote sensing and machine learning: A systematic review[J]. Smart Agricultural Technology, 2024. [2]REN S O，HE K M，GIRSHICK R，t al. Faster R-CNN: towards real- time object detection withregion proposal networksJ. Advances in Neural Information Processing Systems,2015, 28: 91-99. [3]Hanwen Kang,Chao Chen. Fruit detection, segmentation and 3D visualisation of environments in apple orchards[J]. Computers and Electronics in Agriculture,2020,171(C). [4]Liu Anwen,Xiang Yang,Li Yajun,Hu Zhengfang, et al. 3D Positioning Method for Pineapple Eyes Based on Multiangle Image Stereo-Matching[J]. Agriculture,2022 [5]LI Xingxu, CHEN Wenbai, WANG Yiqun, et al.Design and experiment of an Automatic cherry tomato harvesting system Based on cascade vision detection [J]. Transac-tions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2023, 39(1): 136-145. [6]Girshick R , Donahue J , Darrell T ,et al. Region-Based Convolutional Networks for Accurate Object Detection and Segmentation [J].IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 38(1):142-158. [7]ZHENG T ,WANG X ,PENG X , et al.Survey of neurocognitive disorder detection methods based on speech, visual, and virtual reality technologies[J].,2024,6(06):421-472. [8]SUN Jun, QIAN Lei, ZHU Weidong, et al. Apple detection in complex orchard environment based on improved RetinaNet[J]. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 2022, 38(15): 314-322. [9]GU Baoxing, Liu Qin, Tian Guangzhao, Wang Haiqing, Li He, Xie Shangjie. Tree trunk recognition and localization based on improved YOLOv3 [J]. Transactions of the Chinese Society of Agricultural Engineering, 202,38(06):122-129.