Detection and Counting of Tassels for Maize Crop Monitoring using Multispectral Images

Crop monitoring is essential to increase its production and to fulfill future food-demand. For maize crops, the tassels' detection is imperative as tassels give knowledge about its pollination, reproduction stages, and yield. In general, the health and growth stages of a crop are monitored with manual vision, which is time-consuming and labor-intensive. Therefore, Unmanned Aerial Vehicle (UAV) is the main focus of research using different sensors for automating crop-health and growth monitoring processes. However, the identification of objects like tassels is challenging due to the size and number of objects, backgrounds, illuminations, and occlusions of objects in images acquired in UAV based remote sensing. This manuscript proposed a novel method to detect tassels of maize crops using UAV based remote sensing. The proposed method used images of five bands: Red, Blue, Green, Rededge, and Near-InfraRed (NIR). The performance analysis demonstrated that the proposed method detected tassels accurately and outperforms the baseline methods. It also aids advantage in the process of automation of agriculture crop-monitoring.