Proposal and Evaluation of an Automatic Classification Method to Agricultural Work Images Taken by a Network Camera with an Infrared Sensor

Abstract

We have installed network-connected cameras and field servers in an agricultural field. In addition, we have developed a field monitoring system that enables elementary school children to observe the agricultural land from their classroom through the Internet and used it in primary school since 2011. However, a teacher could not use agricultural work images in her class because the number of the images taken by the network cameras with an infrared sensor exceeds 50,000 per year. In this study, we developed a function to add metadata to the sensor detection images for improving easiness of search of the agricultural work images from all images. This paper describes a method for automatically classifying images of agricultural work employing both image processing technology and information on the date and time of image capture.