A Wheat Grade Identification and Bad Grain Detection System Based on EasyDL

DOI: https://doi.org/10.62517/jike.202304310

Author(s)

Hanruyi Liu, Yushuo Bai, Jiaxuan Cao, Yinpeng Li, Jiayue Liu

Affiliation(s)

Hebei University of Science and Technology, Shijiazhuang, Hebei, China

Abstract

At present, there are many problems of insufficient professional testing personnel, low testing equipment and low accurate identification. The wheat grade identification and bad grain testing system based on EasyDL will bring profound changes to the agricultural product quality testing industry. This system is implemented in the AI development platform of Baidu fly pulp EasyDL, and the innovative image recognition technology is used in wheat detection. The image features are extracted through the convolutional neural network, and the supervised learning algorithm is used for model training. Using Requests, OpenCV, Xlsx Writer third-party toolkit for data processing, it can realize the detection of seven types of wheat grains and bud gains, and the detection accuracy can reach 95.94%, and the accuracy of quantitative statistics is 98.95%.

Keywords

Quality and Safety Detection of Agricultural Products; Wheat Detection; Image Recognition Technology; Convolution Neural Network for Image Extraction; Internet+ Crop Quality Detection

References

[1] Zhou Feiyan, J in Linpeng, Dong Jun. Review of convolutional neural network research. Journal of Computer Science, 2017,40 (6) :23. [2] Zhang Yurong, Wang Qiangqiang, Wu Qiong, et al. Based on the Python-OpenCV Wheat imperfect grain recognition study by image processing technology. Journal of Henan University of Technology: Natural Science Edition, 2021,42 (6): 8-10. [3] Deng Qiang, Jing Yuan, CAI Yanhong, et al. Agricultural product quality and safety testing Analysis of the input and output risks in the organization workflow. Journal of Food Safety and Quality Testing, 2016,7 (11): 6-9. [4] Hou Guohu. Agricultural product quality and safety supervision and testing work of the difficulties and Strategy for inquiry. Seed Technology, 2020,38 (5): 12-15. [5] Green rice and my wood. Agricultural product quality safety detection exists in the work and its countermeasures. Southern Agriculture, 2020,14 (5): 2. [6] Feng Xiaoxia. Research on the image recognition algorithm based on deep learning. North China University of Technology,2021. [7] Gao Junyu, Yang Xiaoshan, Zhang Tianzhu, Xu Chang sheng. Based on deep learning Robustness of the visual tracking method. Journal of Computer Science.2016 (07):56-72. [8] Gao Junyu, Yang Xiaoshan, ZhangTianzhu, Xu Chang sheng. Based on deep learning Robustness of the visual tracking method. Journal of Computer Science.2016 (07) [9] Wang Haichao, Kaohsiung, Chen Tiey ing, et al. Crop disease detection Status and prospects of spectroscop ic and image processing technolog ies. Agricultural Research, 2015,37 (10): 69-72. [10] Li Yandong. Key technique of computer vision based on convolutional neural network Surg ical study. University of Electronic Science and Technology of China,2019. [11] Li Yandong, Hao Zongbo, Lei Hang. Review of convolutional neural network research. Computer App lications, 2016,36 (009) :2508-2515. [12] Bradsky, K ler, Yu Shiq i, et al.study OpenCV: Chinese version. Tsinghua University Press, 2020.