Fast recognition method for betel nut in dense environments based on improved YOLO
2023
DAI Yun | LU Ming | HE Ting | PENG Cheng-wu
Objective: This paper aimed to improve the accuracy of identification of small individual betel nuts and the degree of automation of betel nut processing plant by combining with deep learning. Methods: In this study, a novel feature extraction network named Mob-darknet-52 was proposed to construct a method of betel nut location and recognition based on improved YOLO algorithm by using multi-scale detection size. Results: the test showed that the proposed method had a detection accuracy of 94.8%, an accuracy rate of 94.5%, a recall rate of 95.1%, and a detection time of 6.679 ms in betel nut classification. Conclusion: The optimized algorithm based on improved YOLOV3 network can realize the rapid location and identification of betel nut in dense environment.
Show more [+] Less [-]Bibliographic information
This bibliographic record has been provided by Directory of Open Access Journals