Abstract
To address the pollination strategy issues caused by the morphological differences of kiwifruit flowers at different blooming stages and the limited computing power of edge devices, this study proposes a lightweight kiwifruit flower detection method based on the improved YOLOv11n. The method accurately identifies flowers with different opening degrees, and compared with the baseline model, the model performs better on edge computing devices. Specifically, the StarNet lightweight backbone was employed to replace the feature extraction network of YOLOv11n, thereby reducing model complexity; an RGSE module integrating multibranch convolution and attention mechanisms was designed to enhance multiscale feature aggregation; the detection head was redesigned into a Detect_Efficient structure to improve inference efficiency; and the loss function was replaced with MPDIoU. Experimental results demonstrate that the improved model, YOLOv11n‐SREM, achieves a precision of 93.0%, recall of 92.6%, and mean average precision (mAP) of 95.3%. The number of parameters, model weight, and computational cost are reduced to 1.5 M, 3.3 M, and 3.4 GFLOPs, representing decreases of 42.3%, 41.1%, and 46.0% compared to the baseline YOLOv11n, respectively. The model achieves 303.1 FPS, with 24 ms latency and 90.1% detection accuracy on an edge device. These findings provide an effective technical solution for the lightweight deployment of automated kiwifruit pollination systems.
Abstract from the paper’s OpenAlex record, shown for indexing and research. Copyright remains with the authors or the publisher. Read at the source ↗
The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.
Claims 10
Claim · model-written summary
The improved model identified flowers at different opening degrees.
- Taxon
- Kiwifruit
Actinidia deliciosat0 - Outcome
- identifies flowers with different opening degreese0
Not reportedComparator
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“The method accurately identifies flowers with different opening degrees”
t0“Kiwifruit Flower”
Claim · model-written summary
The model achieved 93.0% precision.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- precisione0
- Value
- 93.0%e0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“achieves a precision of 93.0%”
t0“Kiwifruit Flower”
Claim · model-written summary
The model achieved 92.6% recall.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- recalle0
- Value
- 92.6%e0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“recall of 92.6%”
t0“Kiwifruit Flower”
Claim · model-written summary
The model achieved 95.3% mean average precision.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- mean average precision (mAP)e0
- Value
- 95.3%e0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“mean average precision (mAP) of 95.3%”
t0“Kiwifruit Flower”
Claim · model-written summary
Parameter count fell 42.3% to 1.5 M versus baseline.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- number of parameterse0
- Direction
- reducede0
- Value
- 1.5 Me0
- Qualifier
- 42.3%e0
- Comparator
- baseline YOLOv11ne0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“The number of parameters, model weight, and computational cost are reduced to 1.5 M, 3.3 M, and 3.4 GFLOPs, representing decreases of 42.3%, 41.1%, and 46.0% compared to the baseline YOLOv11n, respectively”
t0“Kiwifruit Flower”
Claim · model-written summary
Model weight fell 41.1% to 3.3 M versus baseline.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- model weighte0
- Direction
- reducede0
- Value
- 3.3 Me0
- Qualifier
- 41.1%e0
- Comparator
- baseline YOLOv11ne0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“The number of parameters, model weight, and computational cost are reduced to 1.5 M, 3.3 M, and 3.4 GFLOPs, representing decreases of 42.3%, 41.1%, and 46.0% compared to the baseline YOLOv11n, respectively”
t0“Kiwifruit Flower”
Claim · model-written summary
Computational cost fell 46.0% to 3.4 GFLOPs versus baseline.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- computational coste0
- Direction
- reducede0
- Value
- 3.4 GFLOPse0
- Qualifier
- 46.0%e0
- Comparator
- baseline YOLOv11ne0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“The number of parameters, model weight, and computational cost are reduced to 1.5 M, 3.3 M, and 3.4 GFLOPs, representing decreases of 42.3%, 41.1%, and 46.0% compared to the baseline YOLOv11n, respectively”
t0“Kiwifruit Flower”
Claim · model-written summary
The model achieved 303.1 FPS.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- FPSe0
- Value
- 303.1 FPSe0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“The model achieves 303.1 FPS”
t0“Kiwifruit Flower”
Claim · model-written summary
Edge device latency was 24 ms.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- latencye0
- Value
- 24 mse0
- Setting
- edge devicee0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“with 24 ms latency and 90.1% detection accuracy on an edge device”
t0“Kiwifruit Flower”
Claim · model-written summary
Edge device detection accuracy was 90.1%.
- Taxon
- KiwifruitActinidia deliciosat0
- Outcome
- detection accuracye0
- Value
- 90.1%e0
- Setting
- edge devicee0
Organisms pictured · Representative pictures of the matched species, not figures from the paper.
Verbatim
e0“with 24 ms latency and 90.1% detection accuracy on an edge device”
t0“Kiwifruit Flower”