Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration

Yilun Wen

doi:10.47297/taposatWSP2633-456902.20260703

PDF
导出
分享
收藏
专辑

Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration

7卷第3期

浏览量: 0 Downloads: 3

Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration
Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration
University of Shanghai for Science and Technology, Shanghai, 200093, China
2026年7卷第3期页码：10-15
纸质出版：2026-03-30
DOI：10.47297/taposatWSP2633-456902.20260703
Accepted：

Scan QR Code

Yilun Wen. (2026). Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration. 科技理论与实践（英文版）, 7(3), 10-15.

Wen Yilun.Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration[J].Theory and Practice of Science and Technology,2026,07(03):10-15.
Yilun Wen. (2026). Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration. 科技理论与实践（英文版）, 7(3), 10-15. DOI： 10.47297/taposatWSP2633-456902.20260703.

Wen Yilun.Design of a Strawberry Ripeness Real-Time Detection System Based on YOLOv5-FPGA Acceleration[J].Theory and Practice of Science and Technology,2026,07(03):10-15. DOI： 10.47297/taposatWSP2633-456902.20260703.

摘要

Abstract

To address the problems of low efficiency and poor consistency in traditional strawberry ripeness assessment based on manual observation

a real-time strawberry ripeness detection system based on You Only Look Once version 5 (YOLOv5) and Field Programmable Gate Array (FPGA) acceleration is proposed in this paper. First

a strawberry ripeness object detection dataset is constructed

in which strawberries are classified into three categories: raw

turning

and ripe. Under the same training conditions

two lightweight detection models

YOLOv5n and YOLOv8n are compared and analyzed. Considering both detection performance and hardware implementation complexity

YOLOv5n is finally selected as the base model of the system. Then

with the Xilinx Kintex UltraScale+ chip xcku5p-2ffvb676I as the hardware platform

a Neural Processing Unit (NPU) acceleration architecture supporting operators such as conv

focus

cat

add

Spatial Pyramid Pooling Fast (SPPF)

upsample

reg

and switch is designed. Combined with Advanced eXtensible Interface (AXI) bus transmission

Xilinx Direct Memory Access (XDMA) communication

finite-state-machine control

ping-pong buffering

and pipeline processing

efficient model inference is realized. Experimental results show that the inference time of a single image on the software platform is 74.47 ms

corresponding to 13.43 frames per second (FPS). On the FPGA platform

when the system clock frequency is 200 MHz

the computation time is reduced to 34.3 ms. The results demonstrate that the FPGA-based neural network accelerator can effectively improve the real-time performance of the strawberry ripeness detection system and has promising engineering application value.

关键词

Keywords

references

PANG F, CHEN X. MS-YOLOv5: a lightweight algorithm for strawberry ripeness detection based on deep learning[J]. Systems Science & Control Engineering, 2023, 11(1):2285292.

CHEN Y, XU H, et al. CES-YOLOv8: Strawberry Maturity Detection Based on the Improved YOLOv8[J]. Agronomy, 2024, 14(7):1353.

TAMRAKAR N, KARKI S, et al. Lightweight Improved YOLOv5s-CGhostnet for Detection of Strawberry Maturity Levels and Counting[J]. AgriEngineering, 2024, 6(2):962-978.

MI Z, YAN W Q. Strawberry Ripeness Detection Using Deep Learning Models[J]. Big Data and Cognitive Computing, 2024, 8(8):92.

YUE Y, XU S, WU H. A Strawberry Ripeness Detection Method Based on Improved YOLOv8[J]. Applied Sciences, 2025, 15(11):6324.

LI Z, WANG J, et al. SGSNet: a lightweight deep learning model for strawberry growth stage detection[J]. Frontiers in Plant Science, 2024, 15:1491706.

FU H, LI X, et al. LBS-YOLO: a lightweight model for strawberry ripeness detection[J]. Frontiers in Plant Science, 2025, 16:1715263.

XIAO B, NGUYEN M, et al. Fruit ripeness identification using YOLOv8 model[J]. Multimedia Tools and Applications, 2024, 83(9):28039-28056.

YAN Z, ZHANG B, WANG D. An FPGA-Based YOLOv5 Accelerator for Real-Time Industrial Vision Applications[J]. Micromachines, 2024, 15(9):1164.

SAIDANI T, GHODHBANI R, et al. Hardware Acceleration for Object Detection using YOLOv5 Deep Learning Algorithm on Xilinx Zynq FPGA Platform[J]. Engineering, Technology & Applied Science Research, 2024, 14(1):13066-13071.

QIAN W, ZHU Z, et al. FPGA-based accelerator for YOLOv5 object detection with optimized computation and data access for edge deployment[J]. Parallel Computing, 2025, 124:103138.

HENG C W, UTTRAPHAN C, et al. Optimizing FPGA-based YOLO series accelerators: A survey of techniques[J]. Neurocomputing, 2025, 650:130874.

LIU F, LI H, et al. Review of neural network model acceleration techniques based on FPGA platforms[J]. Neurocomputing, 2024, 610:128511.

WANG D, WANG X, et al. Strawberry ripeness classification method in facility environment based on red color ratio of fruit rind[J]. Computers and Electronics in Agriculture, 2023, 214:108313.