A Comparison of Effectiveness Between 2-Dimensional and 3-Dimensional Data Visualization in Detecting Plant Architectural Characteristics

Abstract

A new generation of crops that yield more with fewer inputs and are adapted to more variable environments is needed to secure food security. The key to breeding highly productive and tolerant crops will be the to analyze the effect of the environment on plant growth through analyzing diverse plant measurements. Plant architecture is an essential variable in plant adaptation to environment and can bring important information to understand physiological processes governing plant functioning. The factors include various images and visualized data such as leaf areas, stem diameters, plant heights and widths, intermodal lengths, color, etc. However, the development of efficient tools assessing plant architecture is still a major stake. Nowadays, due to the advance of image capturing technology, a broad range of image data and quantitative measurements can be obtained. This study explored the relative advantages of 2D versus 3D data visualization for agricultural plant architecture analysis. We compared the effectiveness and efficiency of information processing between 2D and 3D visualization by a human subject experiment. In this study, an automatic feature extraction system (AFES) was developed using the algorithms that can automatically extract plant characteristics from captured plat images. 3D plant models are constructed by height and volume measurements, leaf detection and green index. The extracted features were prepared in 2D or 3D data visualization formats (images or 3D models) for determining plant structure characteristics. The task completion times and accuracy rates were measured as performance indicators in comparison of two data visualization formats. The effect of different data visualization on human information perception were evaluated, particularly in an agricultural data analysis process.