Evaluation amount of rice grains in photograph using image processing techniques

Abstract

Threshing the rice from using combine harvesters has direct impacts variously on the rice yield. The problem caused from the machine itself does not a real matter since the users can manage the machine independently. However, the great unmanageable problem to be primarily considered is the condition of rice in the field such as its growing intensity. Due to the partially inconsistent number of the rice growing in a field, a steady speed of reel index of the combine harvesters will damage the grain in a particular part of the rice field. In the study, the researcher introduces rice grain intensity estimating method by using rice grain screening algorithm system by capturing a digital picture through a camera attached at the top of the combine harvesters in order to further develop the rice combine harvester header with a real-time reel index speed adjusting system. Grains, leaves, stalks and soil surface condition are captured and analyzed concerning the rice moisture content and the altitude at which the picture is captured in real time. The developed rice grain screening algorithm system uses model random forest classification to estimate the rice grain quantity by indicating different color of each part of the rice. During the test, the analyzed result from using rice grain screening algorithm system precisely classified each part of the rice between 59.43 to 86.94 percent of the entire photo pixels. After the process of analyzing, the analyzed intensity of rice production ranges between 3.08 to 6.92 percent. Then the readers can use the result in this study for citation to develop the rice combine harvester header with a real-time reel index speed adjusting system.