A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model

Accurate identification of picking points remains a critical challenge for automated marigold harvesting, primarily due to complex backgrounds and significant pose variations of the flowers. To overcome this challenge, this study proposes SCS-YOLO-Seg, a novel method based on a lightweight segmentat...

Full description

Saved in:
Bibliographic Details
Main Authors: Baojian Ma, Zhenghao Wu, Yun Ge, Bangbang Chen, He Zhang, Hao Xia, Dongyun Wang
Format: Article
Language:English
Published: MDPI AG 2025-08-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/15/4820
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849239709877796864
author Baojian Ma
Zhenghao Wu
Yun Ge
Bangbang Chen
He Zhang
Hao Xia
Dongyun Wang
author_facet Baojian Ma
Zhenghao Wu
Yun Ge
Bangbang Chen
He Zhang
Hao Xia
Dongyun Wang
author_sort Baojian Ma
collection DOAJ
description Accurate identification of picking points remains a critical challenge for automated marigold harvesting, primarily due to complex backgrounds and significant pose variations of the flowers. To overcome this challenge, this study proposes SCS-YOLO-Seg, a novel method based on a lightweight segmentation model. The approach enhances the baseline YOLOv8n-seg architecture by replacing its backbone with StarNet and introducing C2f-Star, a novel lightweight feature extraction module. These modifications achieve substantial model compression, significantly reducing the model size, parameter count, and computational complexity (GFLOPs). Segmentation efficiency is further optimized through a dual-path collaborative architecture (Seg-Marigold head). Following mask extraction, picking points are determined by intersecting the optimized elliptical mask fitting results with the stem skeleton. Experimental results demonstrate that SCS-YOLO-Seg effectively balances model compression with segmentation performance. Compared to YOLOv8n-seg, it maintains high accuracy while significantly reducing resource requirements, achieving a picking point identification accuracy of 93.36% with an average inference time of 28.66 ms per image. This work provides a robust and efficient solution for vision systems in automated marigold harvesting.
format Article
id doaj-art-1dd5d282005a49509ce32ea8cd1fe6f9
institution Kabale University
issn 1424-8220
language English
publishDate 2025-08-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-1dd5d282005a49509ce32ea8cd1fe6f92025-08-20T04:00:51ZengMDPI AGSensors1424-82202025-08-012515482010.3390/s25154820A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg ModelBaojian Ma0Zhenghao Wu1Yun Ge2Bangbang Chen3He Zhang4Hao Xia5Dongyun Wang6Department of Mechanical and Electrical Engineering, Xinjiang Institute of Technology, Aksu 843100, ChinaCollege of Mechanical and Electrical Engineering, Shihezi University, Shihezi 832003, ChinaCollege of Mechanical and Electrical Engineering, Shihezi University, Shihezi 832003, ChinaDepartment of Mechanical and Electrical Engineering, Xinjiang Institute of Technology, Aksu 843100, ChinaCollege of Mechanical and Electrical Engineering, Shihezi University, Shihezi 832003, ChinaCollege of Mechanical and Electrical Engineering, Shihezi University, Shihezi 832003, ChinaCollege of Engineering, Zhejiang Normal University, Jinhua 321004, ChinaAccurate identification of picking points remains a critical challenge for automated marigold harvesting, primarily due to complex backgrounds and significant pose variations of the flowers. To overcome this challenge, this study proposes SCS-YOLO-Seg, a novel method based on a lightweight segmentation model. The approach enhances the baseline YOLOv8n-seg architecture by replacing its backbone with StarNet and introducing C2f-Star, a novel lightweight feature extraction module. These modifications achieve substantial model compression, significantly reducing the model size, parameter count, and computational complexity (GFLOPs). Segmentation efficiency is further optimized through a dual-path collaborative architecture (Seg-Marigold head). Following mask extraction, picking points are determined by intersecting the optimized elliptical mask fitting results with the stem skeleton. Experimental results demonstrate that SCS-YOLO-Seg effectively balances model compression with segmentation performance. Compared to YOLOv8n-seg, it maintains high accuracy while significantly reducing resource requirements, achieving a picking point identification accuracy of 93.36% with an average inference time of 28.66 ms per image. This work provides a robust and efficient solution for vision systems in automated marigold harvesting.https://www.mdpi.com/1424-8220/25/15/4820marigoldsegmentationlightweight modelpicking point recognitionautomated harvesting
spellingShingle Baojian Ma
Zhenghao Wu
Yun Ge
Bangbang Chen
He Zhang
Hao Xia
Dongyun Wang
A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model
Sensors
marigold
segmentation
lightweight model
picking point recognition
automated harvesting
title A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model
title_full A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model
title_fullStr A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model
title_full_unstemmed A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model
title_short A Recognition Method for Marigold Picking Points Based on the Lightweight SCS-YOLO-Seg Model
title_sort recognition method for marigold picking points based on the lightweight scs yolo seg model
topic marigold
segmentation
lightweight model
picking point recognition
automated harvesting
url https://www.mdpi.com/1424-8220/25/15/4820
work_keys_str_mv AT baojianma arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT zhenghaowu arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT yunge arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT bangbangchen arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT hezhang arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT haoxia arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT dongyunwang arecognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT baojianma recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT zhenghaowu recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT yunge recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT bangbangchen recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT hezhang recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT haoxia recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel
AT dongyunwang recognitionmethodformarigoldpickingpointsbasedonthelightweightscsyolosegmodel