Grape Downy Mildew On-line Detection Based on Accurate Image Processing Method Preformed on Embedded Ambient Intelligence System
In this paper, an accurate and intelligent Grape Downy Mildew (GDM) detection method based on common image processing and artificial neural network (ANNs) is proposed. In view of the structure of grape leaves and the distribution characteristics of GDM on leaves, firstly, image processing is mainly to solve the extraction of downy mildew areas on leaves, which is mainly composed of gray-scale processing, gray-scale linear transformation and binarization, and the effects of different processing algorithms and parameters on the results are discussed. Secondly, an ANNs is used to reduce the interference of grape leaf vein on the spot area statistics. Finally, the hardware and software platform based on ARM9 integrated processor, Linux and QT is used to realize the system integration. Compared with the traditional detection method, the accuracy of this detection method can reach 97%, which is close to the accuracy of the human eye, which can be used as an ambient intelligence system in the vineyard inspection, automatic completion of grape GDM detection and grading.
KeywordsGrape Downy Mildew (GDM) Image processing Artificial Neural Network (ANNs) Embedded ambient intelligence system
This work was financially supported by the Chinese National Natural Science Foundation (Grant No. 61673195), The University “Blue_Cyanine project” training plan of Jiangsu, Chinese National Natural Science Foundation (Grant No. 31701324), Science and Technology projects plan of Jiangsu Vocational College of Agriculture and Forestry (Grant No. 2018kj11 and 2018kj12).
- 4.Rupil, C.: Development of image processing and its applications on cryptography. Int. J. Eng. Sci. Res. Technol. 4(7), 1160–1162 (2015)Google Scholar
- 12.Meng, T., Shyu, M.L.: Biological image temporal stage classification via multi-layer model collaboration. In: IEEE International Symposium on Multimedia, pp. 30–37 (2013)Google Scholar
- 13.Yasutake, T.: U.S. Patent No. 5,483,261. U.S. Patent and Trademark Office, Washington, DC (1996)Google Scholar
- 17.Schowengerdt, R.A.: Techniques for Image Processing and Classifications in Remote Sensing. Academic Press (2012)Google Scholar
- 21.Vijay Babu, M.: Real-time object detection based on ARM9. Int. J. Eng. Trends Technol. 4(9), 4080–4083 (2013)Google Scholar
- 23.Liu, J.Z., Li, W.Q., Tian, Y.P.: Automatic thresholding of gray-level pictures using two-dimension Otsu method. In: 1991 International Conference on Circuits and Systems, China. IEEE (1991)Google Scholar
- 26.Zhu, S.P., Xu, H.R., Ying, Y.B., Jiang, H.Y.: Imaging processing technique to measure plant infection severity. In: Proceedings of the SPIE - The International Society for Optical Engineering, vol. 6381 (2006)Google Scholar
- 27.Lydia, B., Jumel, S., Picault, H., Domin, C., Lebreton, L., Ribulé, A., Delourme, R.: An easy, rapid and accurate method to quantify plant disease severity: application to phoma stem canker leaf spots. Eur. J. Plant Pathol. 145(3), 1–13 (2015)Google Scholar
- 28.Kole, D.K., Ghosh, A., Mitra, S.: Detection of downy mildew disease present in the grape leaves based on fuzzy set theory. In: Advanced Computing, Networking and Informatics, vol. 1, pp. 377–384. Springer (2014)Google Scholar
- 29.Chen, Z.L., Zhang, C.L., Shen, W.Z., Chen, X.X.: Grading method of leaf spot disease based on image processing. J. Agric. Mech. Res. 11, 73–75 (2008)Google Scholar
- 30.Tian, Y.W., Chen, X.: The crops leaf disease grade system based on embedded system. J. Shenyang Agric. Univ. 45(6), 756–760 (2014)Google Scholar