We aimed to develop a chatbot application to check seasonal variations in the emergence of the oriental tea tortrix, Homona magnanima Diakonoff (Lepidoptera: Tortricidae) and the optimal timing of insecticide application for this species. The emergence was estimated via image processing. To develop an image-processing program to extract the oriental tea tortrix from the image of the sticky board, we first examined the RGB and HSV color spaces. We found that converting the HSV color space into a black and white image was effective for the model. Thereafter, we determined the optimal shooting time of the web camera by identifying the periods during which the oriental tea tortrix was not captured by the sticky board placed on the canopy surface. Although there were seasonal variations, oriental tea tortrix was not captured on the sticky board between AM4:00 and before sunrise. Thus, it can be speculated that the time period was optimal for image processing. Finally, we developed the program and linked it to a LINE app used by many tea farmers. In summary, we developed a user-friendly chatbot application that can be accessed any time without any complicated operations. This application enhances both accuracy and usability of insecticides against oriental tea tortrix infestation.

In Shizuoka Prefecture, a simple method to monitor the growth of tea plants for tea production is required because of the increase in the managed area per producer. In this study, we developed a technology for estimating the tea yield of plucked new shoots and their composition using a GreenSeeker 2 (GS), an inexpensive, portable NDVI measuring device.

The GS measurement conditions and the relationship between GS-NDVI and yield and between GS-NDVI and composition of the main variety 'Yabukita' were evaluated. In addition, estimation accuracy of 'Yabukita' and other varieties were verified; verification of errors among two devices of the same model was also conducted. The results showed that GS-NDVI was useful for estimating the yield of the first crop of 'Yabukita', because GS-NDVI was highly correlated with the yield. However, GS-NDVI might not be suitable to estimate the second tea crop yield because of the small change. Differences in estimation were observed among varieties other than 'Yabukita', suggesting the need to develop estimation formulas for each variety. In addition, measurement errors were observed between the two devices; however, they could be corrected.

A survey on the vertical distribution on tea hedgerow of the tea harming Mahasena aurea larvae revealed that larval density was highest in the upper layers of the tea trees in mid-May，but the difference between the middle and lower layers gradually decreased. The percentages of removed larvae by light skiffing of canopy (removal of the upper layer) and deep trimming of canopy (removal of the upper and middle layers) in 2017 were 50.8% on June 8^th and 85.4% on July 18^th, respectively. Of the larvae released between the hedges, 18.7% climbed up the tea trees. Larval removal by pruning should be completed by early June before emergence. The highest cause of larval mortality was parasitism by a tachinid Neophryxe psychidis, with parasitism rates were 25.3% on June 8th and 69.6% on July 18^th. The percentages of individuals suspected to have died of disease were 39.7% and 0% on those respective dates. In the future, the development of pest control techniques that combines larval removal by pruning and conservative use of natural enemies are desired.