Early detection of wheat disease based on thermal infrared imaging

Abstract: Wheat can be infected by various pathogens, resulting in a variety of disease symptoms. Early detection of the pathogen infection before symptom appearance is essential for disease monitoring and warning, disease control and the improvement of wheat production. Wheat stripe rust caused by Puccinia striiformis f. sp. tritici (Pst) is an important airborne fungal disease. This study investigated the possibility of thermal infrared imaging technology used for early detection of wheat stripe rust since the Pst infection can result in changes of transpiration rate of wheat leaves and thus may result in changes in leaf temperature. The healthy wheat plants were subjected to artificial spray inoculation with Pst urediospore suspension in an artificial climate chamber. Thermal infrared images and data on leaf temperature of Pst-infected wheat plants in incubation period (no symptom appearance) were collected daily using thermal infrared imaging technology. Meanwhile, healthy wheat plants and diseased wheat plants with symptoms infected with Pst were chosen as controls. The variation of thermal infrared images and leaf temperature of wheat plants was analyzed. The results showed that the infection with Pst resulted in a decrease in temperature at the infected parts of wheat leaves, but the leaves of healthy wheat plants and diseased wheat plants with symptoms had no obvious changes in average temperature and maximum temperature difference (MTD) at different parts of leaves. On 5 days post inoculation (dpi), the positions of the leaves of Pst-infected wheat plants in incubation period in the thermal infrared images appeared light yellow and these Pst-infected wheat plants could be distinguished from the healthy plants by visual observation of thermal infrared images. On 8 dpi, there were yellow spots appearing in the thermal infrared images of Pst-infected wheat plants in incubation period. Then on 12 dpi, the yellow spots turned green in the thermal infrared images and there were small uredinia appearing on the wheat leaves. Data analysis of leaf temperature showed that on 3 dpi, there were significant differences (P<0.05) between the leaves of Pst-infected wheat plants in incubation period and the leaves of healthy wheat plants in average temperature and MTD, and that the leaves of Pst-infected wheat plants exhibited a decease in average temperature by 0.08℃, lower than the leaves of healthy wheat plants, but exhibited an increase in MTD by 0.04℃, higher than the leaves of healthy wheat plants. So wheat plants in incubation period without disease symptoms could be distinguished from the healthy wheat plants via temperature difference. With the increase of the number of days after inoculation, the difference in average temperature and MTD between the leaves of Pst-infected wheat plants and the leaves of healthy wheat plants gradually increased. On 12 dpi, the average temperature of the leaves of Pst-infected wheat plants decreased by 1.22℃ than that of the leaves of healthy wheat plants, and MTD of the leaves of Pst-infected wheat plants increased by 1.58℃ than that of the leaves of healthy wheat plants. Meanwhile, with the increase of the number of days after inoculation, the difference in average temperature and MTD between the leaves of Pst-infected wheat plants in incubation period and the leaves of diseased wheat plants with symptoms gradually decreased. And on 12 dpi, the average temperature of the leaves of Pst-infected wheat plants exhibited 0.27℃ higher than that of the leaves of diseased wheat controls, and MTD of the leaves of Pst-infected wheat plants was 0.30℃ lower than that of the leaves of diseased wheat controls. The results indicated that Pst infection could be detected by visual observation of thermal infrared images at 7 days before the disease symptom apparently appeared and that the infection could be detected by analyzing temperature difference based on thermal infrared images at 9 days before the disease symptom was observed. The results demonstrated that early detection of wheat stripe rust can be realized using thermal infrared imaging technology. This study provides a new method based on thermal infrared imaging for early detection of wheat diseases.