Mobile Application Development for Spectral Signature of Weed Species in Rice Farming

Abstract

Weed infestation happens when there is intense competition between rice and weeds for light, nutrients and water. These conditions need to be monitored and controlled to lower the growth of weeds as they affected crops production. The characteristics of weeds and rice are challenging to differentiate macroscopically. However, information can be acquired using a spectral signature graph. Hence, this study emphasises using the spectral signature of weed species and rice in a rice field. The study aims to generate a spectral signature graph of weeds in rice fields and develop a mobile application for the spectral signature of weeds. Six weeds were identified in Ladang Merdeka using Fieldspec HandHeld 2 Spectroradiometer. All the spectral signatures were stored in a spectral database using Apps Master Builder, viewed using smartphones. The results from the spectral signature graph show that the jungle rice (Echinochloa spp.) has the highest near-infrared (NIR) reflectance. In contrast, the saromacca grass (Ischaemum rugosum) shows the lowest NIR reflectance. Then, the first derivative (FD) analysis was run to visualise the separation of each species, and the 710 nm to 750 nm region shows the highest separation. It shows that the weed species can be identified using spectral signature by FD analysis with accurate separation. The mobile application was developed to provide information about the weeds and control methods to the users. Users can access information regarding weeds and take action based on the recommendations of the mobile application.