Sjukdomsidentifiering i potatis och stråsäd med luftburna multispektrala sensorer

Abstract: Potato (Solanum tuberosum L.) and winter wheat (Triticum aestivum L.) fields with ongoing field trials testing use of fungicides were evaluated using two camer-as (RGB-sensors that registers light in the red (R), green (G) and blue (B) wave-length bands) and one multispectral sensor (that registers light in five narrow bands in the visible and near infrared region of the electromagnetic spectrum). Data was collected using the three sensors at multiple times throughout the season, Seven times from the 20th of June to August 17th in parcels with potatoes and five times from 8th of June to July 23rd in trials with winter wheat. The collected data were later used in creating a mosaic, from which values of reflectance was extract-ed and compared to traditional (manual) methods of estimating the severity of diseases. This was carried out throughout the season. The ambition of the present thesis was to test whether it is possible to use Unmanned Aerial Vehicle (UAV)-borne sensors to detect diseases in fields with potatoes and winter wheat. The evaluated diseases were potato late blight (Phytophtora infestans) and septoria leaf blotch (Septoria tritici). In winter wheat we were not able to find any correlation neither between single bands of reflectance nor vegetation indices (GRVI) and the amount of S.tritici in the crop (p>0.05). However, the GRVI-index was found suitable to evaluate the amount of green leaf area (p= 0.05) in the canopy. The season in which this study was carried out (2016) lacked in rain during June and July which might have contributed to the unusually low amount of fungal diseases recorded. The present thesis concludes that it is possible to detect areas with pos-sible infections due to changes in values of reflectance using multispectral sensors and its correlation to infections. It is possible to use it, together with manual field observations, as decision support for application of fungicides. However it is not certain the disease is detected early enough not to spread further into the field.