Fecha de publicación: 18/03/2021 Fuente: WIPO Agriculture Portada The major problem faced by today's world is water scarcity. If the effective usage of water is minimal when compared with the total amount of water drawn, then water wastage is more. Even though the water scarcity persists, agriculture have to be sustained in order to satisfy the food requirements of human race. With the help of recent technologies, we can design a smart irrigation model which in turn maximizes the amount of effective usage of water in agriculture field.
In this work, three varieties of sensors are used such as climatic sensors, soil sensors and soil NPK sensors. The climate sensors assess the environmental conditions and gathers the details like temperature, humidity and wind speed. These are the main factors which are responsible for rainfall. If there is a possibility for excess rain or excess wind, then the irrigation to the agriculture field will be delayed to certain time. This feature of our model is developed with deep learning technique, which in turn improves the efficiency of the machine by learning from the previous results. The soil and NPK sensors gathers the details like moisture, NPK nutrients traces in the soil of the agriculture field. The input fed to this system is, soil type, relative humidity and plant type. The output will be suggestion predicting the need for irrigation with a period and also the amount of fertilizer needed to provide the maximum yield in an agriculture field.
The radio circuit model gathers the details from climate sensors and soil sensors. The data measurements of the parameters such as temperature, humidity and wind speed will be given as input to the input layer of the Convolutional Neural Network (CNN). The processing will take place in the hidden layer and the following outputs will be produced.
• Due to the prediction of excess wind in the near future, the distribution of water in unnecessary area will be avoided.
• The system could provide a suggestion regarding the irrigation at right time and right quantity.
• The duration of irrigation for the agriculture field is predicted accurately with the help of moisture sensors.

Accurate mixing proportion offertilizer as per the NPK traces available in the soil surface Convolutional . Neura Estimating the right quantity of irrigation at the right time Suggestions Avoiding wsatage of water by predicting the near future environmental conditions

Figure 1. Smart irrigation by incorporating data from climate, soil and NPK sensors and deep learning techniques