PRINT ISSN 2285-5653, CD-ROM ISSN 2285-5661, ONLINE ISSN 2286-1580, ISSN-L 2285-5653


Published in Scientific Papers. Series B, Horticulture, Vol. LXIII, Issue 1
Written by Rodica SOBOLU, Mirela CORDEA, Ioana POP, Daniela POPESCU, Dana PUSTA

Automatic detection of vineyards’ Downey Mildew based on image processing techniques provides better results compare to visual observation performed by farmers. This technique can detect leaf attack even in the onset phase and it can prevent the spread of infection throughout the vineyard. In addition, this technique can be implemented into relative giant fields, requires a reduced amount of time, lower costs and identifies the disease fast and accurate. In this work, we monitored the degree of attack of two cultivars of vines, Sauvignon Blanc and Fetească Regală, from the vineyard of UASMV Cluj-Napoca. The images taken from the vineyard were loaded into the Matlab application. In the first stage, leaves were preprocessed with a median filter. In the next step, in order to detect the typical spots, we transformed the images into color spaces: RGB, YcbCr, HSV and CieLab. In these color spaces we applied the image segmentation techniques based on threshold methods. The experimental results obtained show that in HSV model space the disease was quite correctly recognized and Sauvignon Blanc variety was most severely affected.

[Read full article] [Citation]

The publisher is not responsible for the opinions published in the Volume. They represent the authors’ point of view.

© 2019 SCIENTIFIC PAPERS. Series B. Horticulture. All Rights Reserved. To be cited: SCIENTIFIC PAPERS. Series B. Horticulture.

Powered by INTELIDEV