Author：Department of Agronomy, National Chung Hsing University

Rice is the major crop in Taiwan with extensive plantation and its production is closely related to the economy and people's livelihood. In recent years, however, agriculture has suffered more damages by climate changes and extreme weather and is associated with higher risks in the process of crop cultivation and production which makes an impact on the food supply stability. In addition to the climatic and environmental factors, problems also exist in terms of the aging and emigration of rural population and the consequent shortage of farming manpower. The agricultural industry is highly dependent on the climate and ecology, and also labor-intensive. To encourage the involvement of more young farming practitioners, the Department of Agronomy of National Chung Hsing University (NCHU) has developed a new model of rice cultivation, with technologies used to raise the efficiency of field management and the productivity.

Rice production in Taiwan is a highly mechanized and mature industry, with a complete system of seedling, soil preparation, transplanting, harvesting, storage, and packaging. For the rice field management, however, most farmers still rely on experience accumulated and inherited from predecessors to judge the timing and amount of fertilization, irrigation, and pesticide application. A model of intelligent rice cultivation management is in demand, to use intelligent facilities based on technology R&D, to assist farmers in the intelligent management of the rice cultivation process, lower the vulnerability of the industry and optimize the industrial environment, so as to enhance the willingness and confidence of new farming practitioners. The intelligent sensing devices introduced include rice field water level gauges, soil EC detectors, water flow calculation meters, field weather stations and drones, etc. A human-machine collaborated intelligent environment is developed to cope with manpower demand facing the aging rural labor force. Mobile devices are used for real-time detection of rice field conditions (See Figure 1).

In the management of rice field irrigation and fertilization, there are proverbs like: "Irrigation decides the production, drainage decides the productivity" and "No fertilizer, no food; no food, no troops." These indicate the importance of water and fertilizer in rice cultivation management. In 2018, the Department of Agronomy of NCHU collaborated with the Council of Agriculture to start the project of an intelligent cultivation management system for early-maturing japonica rice. Miaoli District Agricultural Research and Extension Station and Taichung District Agricultural Research and Extension Station joined the project in 2019 so the project expanded its scope to three experiment sites. In the project, the intelligent cultivation management system for early-maturity and medium-late maturity japonica rice, and for medium-late maturity indica rice is developed and optimized. Through the introduction of smart sensors, including the field micro-weather stations, water level gauges, remote electronic water meters, soil conductivity meters, etc. the research team is able to instantly transmit field parameters to the database. The field environment and water levels are constantly monitored to decide the switching on and off of the irrigation solenoid valves and to facilitate the field management of rice irrigation and drainage in the most efficient manner.

The results of intelligent water management show that, for the early-maturity japonica rice, water was saved by 30.9% and 49.8% in the first and second cultivations of 2018 respectively by using alternative wet and dry (AWD) management compared with the conventional cultivation method. In 2019, the intelligent water management showed water saving rates of 37.6%, 65.3%, and 57.4% in the three different test sites experimenting with the early-maturity and medium-late maturity japonica rice, and medium-late maturity indica rice. The tests on the physiological properties like plant height, leaf age, tiller number, leaf area, and the amount of yield indicated no significant differences from those obtained by conventional cultivation methods. This confirms the water saving effect without reducing the productivity. The intelligent devices make it possible for farmers to monitor the water use in the rice field from home and control the irrigation water volume through the remote electronic water meter. This is a more efficient use of agricultural water resources, and also serves to promote water-saving rice cultivation.

The traditional nutrient management is mainly by on-site manual examination of the physiological properties of the rice, which is tedious and time-consuming. Nowadays, the technologies of droning, aerial photography, intelligent image analysis, and data transmission are booming, and are extensively applied in agricultural R&D, including serving as innovative uses in field nutrient management. The research team collaborated with the Dept. of Civil Engineering of NCHU in 2019 to exercise drone photography on large areas of farmland, and, at the same time, investigated the physiological properties like the plant height, tiller number, SPAD value of leaf color, chlorophyll content, and plant nitrogen in areas of different nutrient treatments (Figure 2). With the data of plant growth index, green coverage, and plant height obtained from rice field image analysis, correlation analysis is conducted on the parameters of the images and the physiological parameters of rice plant nutritional conditions, and a large amount of information is integrated to serve as a decision-making basis for nutrient management during rice cultivation. Through cross-disciplinary cooperation, the results of image analysis of rice leaves obtained by drone photography are used to correspond with the nutritional status of the plants at their peak tillering stage. This becomes a decision-making mode on the recommended amount of earing fertilizer applied.

The rise of smart sensors and the collection and analysis of big data enhances the timeliness and convenience of water and nutrient management in the rice fields. A huge amount of data is obtainable in the future through intelligent management systems, which, incorporated with expert knowledge and parameter identification, plus the expert decision-making system achieved with AI deep learning, will achieve the goal of intelligent rice field management. The use of smart sensors and drones on rice fields is a convenient and trendy management model. Hopefully this will inspire the involvement of more young farming practitioners, and promise a future of smart applications for rice farmers.