Abstract
Florescence information monitoring is essential for strengthening orchard management activities, such as flower thinning, fruit protection, and pest control. A lightweight object recognition model using cascade fusion YOLOv4-CF is proposed, which recognizes multi-type objects in their natural environments, such as citrus buds, citrus flowers, and gray mold. The proposed model has an excellent representation capability with an improved cascade fusion network and a multi-scale feature fusion block. Moreover, separable deep convolution blocks were employed to enhance object feature information and reduce model computation. Further, channel shuffling was used to address missing recognition in the dense distribution of object groups. Finally, an embedded sensing system for recognizing citrus flowers was designed by quantitatively applying the proposed YOLOv4-CF model to an FPGA platform. The mAP@.5 of citrus buds, citrus flowers, and gray mold obtained on the server using the proposed YOLOv4-CF model was 95.03%, and the model size of YOLOv4-CF + FPGA was 5.96 MB, which was 74.57% less than the YOLOv4-CF model. The FPGA side had a frame rate of 30 FPS; thus, the embedded sensing system could meet the demands of florescence information in real-time monitoring.
Highlights
Florescence information monitoring is the fundamental technical basis and key management chain for achieving a high-yield and high-quality orchard for strengthening orchard management activities, such as flower thinning, fruit protection, pest control, and yield prediction [1,2]
This study is outlined as follows: Section 2 shows the collection and processing of experimental data; Section 3 introduces the design of the YOLOV4-CF model for recognizing citrus flowers and gray mold; Section 4 presents the YOLOV4-CF model migration and deployment process of the field programmable gate array (FPGA) embedded platform; Section 5 analyzes the experimental results in detail; Section 6 concludes the study
The results of the transfer learning training demonstrate that the precision mAP@.5 value of the YOLOv4-CF model is 95.03%, and the F1 value is 89.00%, which is 1.42% higher than that of the YOLOv4-Tiny model, and the recognition precision of the citrus flower, citrus bud, and gray mold is improved by 0.94%, 0.72, and 2.07%, respectively
Summary
Florescence information monitoring is the fundamental technical basis and key management chain for achieving a high-yield and high-quality orchard for strengthening orchard management activities, such as flower thinning, fruit protection, pest control, and yield prediction [1,2]. Research on recognizing bud, flower, and gray mold in a natural environment is in line with the application requirements of citrus florescence information monitoring. YOLOv4-CF, which is a lightweight object recognition model for citrus bud, flower, and gray mold, was proposed using the software and hardware codesign pattern. A lightweight object recognition model using cascade fusion YOLOv4-CF is proposed, which recognizes multi-type objects in their natural environments, such as citrus buds, citrus flowers, and gray mold. 3. An embedded sensing system for recognizing citrus flowers is designed by quantitatively applying the proposed YOLOv4-CF model to an FPGA platform. This study is outlined as follows: Section 2 shows the collection and processing of experimental data; Section 3 introduces the design of the YOLOV4-CF model for recognizing citrus flowers and gray mold; Section 4 presents the YOLOV4-CF model migration and deployment process of the FPGA embedded platform; Section 5 analyzes the experimental results in detail; Section 6 concludes the study
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
