Abstract
[Objective] To solve the problems of low efficiency and high cost in cherry screening. [Methods] A heatmap regression method HRNet-YT was proposed to automatically identify cherry size and the presence of fruit stems, thereby realizing efficient screening. HRNet-YT utilized multiple parallel subnetworks to achieve multi-scale information fusion while maintaining high-resolution representations, ensuring the spatial accuracy of heatmaps for stem and calyx keypoints. By leveraging heatmap techniques to capture rich contextual information and optimizing the loss function, the model's robustness and precision were enhanced. [Results] HRNet-YT-W48 (384×288) achieves a detection precision of 87.3% and an keypoint average precision (AP, OKS =0.5) of 0.22 on the dataset. [Conclusion] The proposed method demonstrates high precision and adaptability in the cherry keypoint detection task.
Publication Date
1-23-2026
First Page
72
Last Page
78
DOI
10.13652/j.spjx.1003.5788.2024.60147
Recommended Citation
Xuejun, SONG; Lei, GAO; and Xiaoxia, GUO
(2026)
"Cherry grading screening based on deep learning-driven heatmap regression,"
Food and Machinery: Vol. 42:
Iss.
1, Article 10.
DOI: 10.13652/j.spjx.1003.5788.2024.60147
Available at:
https://www.ifoodmm.cn/journal/vol42/iss1/10
References
[1] BALAN V,IVANOV I,ŞARBAN V,et al.Changes in cherries size and quality according to color [J].Lucrări Ştiinţifice Seria Horticultură,2017,60(2):139-144.
[2] ROPELEWSKA E,POPIŃSKA W,SABANCI K,et al.Cultivar identification of sweet cherries based on texture parameters determined using image analysis [J].Journal of Food Process Engineering,2021,44(7):e13724.
[3] 王 昭.基 于 机 器 视 觉 的 樱 桃 缺 陷 检 测 [D].北 京:北 京 林 业 大学,2019:18-27.WANG Z.Cherries defect detection based on machine vision[D].Beijing:Beijing Forestry University,2019:18-27.
[4] 裴悦琨,叶家敏,姜艳超,等.基于机器视觉的樱桃形状及大小检测技术 [J].食品工业,2020,41(8):199-202.PEI Y K,YE J M,JIANG Y C,et al.The cherry shape and size detection technology based on machine vision [J].The Food Industry,2020,41(8):199-202.
[5] 杨立伟,赖文聪,刘刚,等.基于改进 Harris 角点检测的虚拟樱桃叶片重建方法 [J].农业机械学报,2022,53(S1):213-217.YANG L W,LAI W C,LIU G,et al.Virtual leaf reconstruction method based on improved Harris corner detection [J].Transactions of the Chinese Society for Agricultural Machinery,2022,53(S1):213-217.
[6] 魏 冉,裴 悦 琨,姜 艳 超,等.基 于 改 进 Faster R-CNN 模 型 的 樱桃缺陷检测 [J].食品与机械,2021,37(10):98-105,201.WEI R,PEI Y K,JIANG Y C,et al.Detection of cherry defects based on improved Faster R-CNN model [J].Food & Machinery,2021,37(10):98-105,201.
[7] 李兴旭,陈雯柏,王一群,等.基于级联视觉检测的樱桃番茄自 动 采 收 系 统 设 计 与 试 验 [J].农 业 工 程 学 报,2023,39(1):136-145.LI X X,CHEN W B,WANG Y Q,et al.Design and experiment of an automatic cherry tomato harvesting system based on cascade vision detection [J].Transactions of the Chinese Society of Agricultural Engineering,2023,39(1):136-145.
[8] 代东南,马睿,刘起,等.基于特征融合注意力机制的樱桃缺陷检测识别研究 [J].山东农业科学,2024,56(3):154-162.DAI D N,MA R,LIU Q,et al.Research on cherry defect detection and recognition based on feature fusion attention mechanism [J].Shandong Agricultural Sciences,2024,56(3):154-162.
[9] 赵博,柳苏纯,张巍朋,等.面向樱桃番茄采摘识别的轻量化Transformer 架 构 优 化 研 究 [J].农 业 机 械 学 报,2024,55(10):62-71,105.ZHAO B,LIU S C,ZHANG W P,et al.Performance optimization of lightweight transformer architecture for cherry tomato picking [J].Transactions of the Chinese Society for Agricultural Machinery,2024,55(10):62-71,105.
[10] 沈超,夏秀波,杨玮,等.温室樱桃番茄果实深度学习识别模型研究 [J].北方农业学报,2023,51(5):114-122.SHEN C,XIA X B,YANG W,et al.Research on deep learning recognition model of greenhouse cherry tomato fruit [J].Journal of Northern Agriculture,2023,51(5):114-122.
[11] QIN X Y,CAO J X,ZHANG Y H,et al.Development of an optimized YOLO-PP-based cherry tomato detection system for autonomous precision harvesting [J].Processes,2025,13(2):353.
[12] DU W S,JIA Z H,SUI S S,et al.Table grape inflorescence detection and clamping point localisation based on channel pruned YOLOV 7-TP[J].Biosystems Engineering,2023,235:100-115.
[13] AMIRI N,WANG G H,JANABI-SHARIFI F.Keypoint detection technique for image-based visual servoing of manipulators [C]// 2024 IEEE 20th International Conference on Automation Science and Engineering (CASE ).Bari,Italy:IEEE,2024:3 557-3 562.
[14] LI Y J,FENG Q C,LIU C,et al.MTA-YOLACT:multitask-aware network on fruit bunch identification for cherry tomato robotic harvesting [J].European Journal of Agronomy,2023,146:126812.
[15] 杨震宇,汪小旵,祁子涵,等.基于改进 YOLOv 8的草莓识别与 果 梗 采 摘 关 键 点 检 测 [J].农 业 工 程 学 报,2024,40(18):167-175.YANG Z Y,WANG X C,QI Z H,et al.Recognition strawberry to detect the key points for peduncle picking using improved YOLOv 8 model [J].Transactions of the Chinese Society of Agricultural Engineering,2024,40(18):167-175.
[16] 刘敬宇,裴悦琨,常志远,等.基于改进 YOLOX 模型的樱桃缺陷及分级检测 [J].食品与机械,2023,39(1):139-145.LIU J Y,PEI Y K,CHANG Z Y,et al.Cherry defect and classification detection based on improved YOLOX model [J].Food & Machinery,2023,39(1):139-145.
[17] 缑新科,常英.基于机器视觉的樱桃番茄在线分级系统设计[J].计算机与数字工程,2020,48(7):1 798-1 803.GOU X K,CHANG Y.Design of cherry tomato online grading system based on machine vision [J].Computer and Digital Engineering,2020,48(7):1 798-1 803.
[18] 张永飞,裴悦琨,姜艳超,等.基于深度卷积神经网络的樱桃分级检测 [J].食品研究与开发,2021,42(14):138-144.ZHANG Y F,PEI Y K,JIANG Y C,et al.Cherry grading using a deep convolutional neural network [J].Food Research and Development,2021,42(14):138-144.
[19] ISLAM M A,JIA S,BRUCE N D B.How much position information do convolutional neural networks encode ? [EB/OL].arXiv preprint arXiv.(2020 -01-22) [2024 -06-11].https://arxiv.org/abs/ 2001.08248 v1.
[20] MAJI D,NAGORI S,MATHEW M,et al.Yolo-pose:enhancing yolo for multi person pose estimation using object keypoint similarity loss [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2022:2 637-2 646.
