Labor-saving weed control technology is essential for expanding organic rice farming. One such technology is an automatic weed suppression robot （referred to as AW） that moves freely in paddy water after transplanting rice seedlings and suppresses weed emergence because of suspended paddy water. Working hours of rice production process were investigated when AW was operated in organic rice paddy in Bando city, Ibaraki Prefecture, and compared to traditional weeding system with mechanical and manual weeding. The AW system reduced the total working hours by up to 30%. In particular, the working hours of mechanical and manual weeding were reduced by 87% and 79%, respectively. However, mechanical weeding was required in the paddy with many weeds, even when AW was operated. In AW system, the time spent plowing, harrowing, leveling, levee coating, and repairing levee increased because the stable operation of AW requires careful implementation of tasks to level the field and maintain water depth. Although the AW system increased the frequency of paddy patrols for water management, the total working hours remained unchanged due to a reduction in the working hours per patrol. The reduced hours per patrol were attributed to careful levee coating before and after transplanting.

A questionnaire survey was conducted on low visibility around agricultural machinery and equipment to improve the low visibility. The survey aimed to clarify issues related to visibility around riding-type agricultural machinery. A total of 479 farmers visited the Institute of Agricultural Machinery, National Agriculture, and Food Research Organization from 2022 to 2024 and answered the questions. Of the respondents, 88.9% recognized the problem of low visibility in agricultural machinery, especially riding tractors with equipment, combine harvesters and rice transplanters. Approximately 48.0% to 71.6% of the respondents were requested to improve their visibility in the locations behind these three machines. Camera monitors, safety sensors, alarm systems, mirrors, AI cameras, and automatic operations are expected to improve low visibility. Available costs for installing such equipment into agricultural machinery were 20,000–50,000 yen for mirrors, 20,000–100,000 yen for camera monitors and automatic operation, 30,000–50,000 yen for safety sensors and alarm systems, and 30,000–100,000 yen for the AI cameras.

Soil moisture is an important factor that determines the workability of agricultural machinery and crop growth. However, existing remote sensing methods by which to estimate soil moisture are adversely affected by both the external environment at the time of imaging and the high costs of sensors. Hence, in this study, we develop a Convolutional Neural Network （CNN） model for soil moisture estimation. The proposed model uses visible light images to estimate soil moisture under wet conditions with high accuracy. Moreover, we evaluate the performance of the proposed model in the establishment of an inexpensive remote sensing method for soil moisture that is less susceptible to external environmental factors. Therefore, the proposed model is an inexpensive, simple, and environmentally independent method by which to estimate soil moisture, compared with existing remote sensing methods.

This study aimed to estimate the damage rate caused by the expansion of foot rot disease in sweet potatoes using the supervised classification of RGB images from an unmanned aerial vehicle （UAV）. We compared the estimation accuracy of three different types of supervised classification models to classify the coverage of sweet potato fields. The model using random forest classification （RFC） achieved the highest accuracy and healthy leaves, yellow leaves, dead leaves, mulches, and bare soils in the field were classified with 0.880 accuracy. The damaged plant rate was estimated using the sum of the rates of yellow and dead leaves. The estimated value was significantly correlated with the value measured from field measurements, indicating that the ‘damaged plant rate’ caused by foot rot disease could be estimated using RFC model. Furthermore, the created model can visualize damage occurrences and expansion. However, the accuracy depletion in estimation of damaged plants was recognized when field was covered with many weeds because the image could not capture the leaf condition of sweet potatoes.