Computer vision techniques that focus on segmenting point clouds have become increasingly popular in agriculture for tasks such as studying plant characteristics and growth patterns of crops in fields overtime using three-dimensional data instead of traditional two-dimensional data can provide more detailed and useful information to scientists. However, due to the scarce availability of point cloud datasets, there are challenges that need to be addressed through improvements in how data is collected and processed This study introduces a method that compares and efficiently separates tapioca plants, from uneven ground surfaces. We obtained real-time data from fields using photogrammetry and then processed it with Cloud Compare software to improve the quality and efficiency of the point cloud data by reconstructing the data and removing noise through down sampling techniques. In order to distinguish tapioca plants, from the soil surface effectively, we utilized the Random Sample Consensus (RANSAC) algorithm that accurately identified and classified the points. Moreover, we utilized RGB alteration to assign colors to individual points simplifying visual inspection and allowing for a clear differentiation, between the tapioca plants and the nearby terrain. The suggested approach showcases efficiency by successfully accomplishing precise segmentation and color assignment of the point cloud data. Consequently, the accurate recognition and categorization of tapioca plants and the adjacent ground surface can be achieved effectively.