Development of Damage Assessment Method of Pests and Diseases of Rice Using multispectral imagery Approach

Information about the level of pests and diseases of rice plants is needed in planning policies related to agricultural production and insurance. The level of pest and disease attacks can be done through several approaches including through a manual (visual) approach, remote sensing technology using satellite, and UAV technology using visible images (Red, Green, Blue). Each of these approaches has advantages and disadvantages. This research was conducted to develop an accurate method for assessing pests and diseases of rice plants using the multispectral image approach. The specific objectives are 1) To determine the correlation between the intensity of pest and disease attacks with several multispectral image parameters, and 2) To establish the equation for estimating the intensity of pest and rice plant attack based on the multispectral images taken. Multispectral image capture is performed using DJI Inspire 1 drones equipped with 5 band multispectral cameras (Red, Green, Blue, Near Infra and Infra Red Band). The image is taken from a height of 40 m and performed at a time interval of 1.5 seconds. Multispectral images obtained at this stage are then analyzed using ArcGis software to analyze the observed variables. The variables observed included the value of NDVI (Normalized Difference Vegetation Index), NDRE (Normalized Difference Red Edge), and the intensity of pest and rice disease attacks. When taking multi-spectral images, measurements of the intensity of disease attacks are also carried out manually according to standard procedures for determining pests and diseases issued by the Ministry of Agriculture. Correlation analysis is then performed to obtain the relationship between the observed multi spectral image variables (NDVI and NDRE) with the intensity of the attack, and to get the equation for estimating the intensity of the attack based on the parameters of the multispectral image. The results showed that the NDVI and NDRE imagery had a strong linear correlation with the intensity of the attacks. It can be concluded that the intensity of rice diseases can be estimated well using the multispectral image approach.