Figure 2 : Insect pests were annotated and labelled by drawing the object bounding boxes on the image using Python language at the object processor .
Furthermore , the same issue of limited pilot assignment in the cellular band is also possible . More research investigating the use of the latest IoT communication standards such as 3GPP NB- IoT and 5G-IoT will enable large-scale pilot tests of I-IoT in agriculture .
Next , energy efficiency of IoT devices remains an issue in agricultural IoT deployment . For precision agriculture , agricultural sensors or IoT devices are spread widely across thousands of acres of farmland . If the devices cannot sustain sufficient battery life , it will affect IoT deployment reliability . Energy efficient schemes for IoT should be further explored , such as the ability to predict the sleep interval of IoT devices based on their remaining battery level , previous usage history , and quality of data required for a particular application .
As agricultural I-IoT applications comprise a combination of several applications ranging from simple rule-based algorithms to advanced computational intelligent processing , the development of agricultural IoT platform will shift towards interoperability capability . The universal or generic platform should be able to support any kind of crop and livestock . Increasing number of work is expected on the use of artificial intelligence such as deep learning for crop productivity estimation
15