Characterization of plant features is essential for effective management of plant growth monitoring, and precise management practices in crop production. Stereo vision captures multiple perspective images to create a threedimensional (3D) representati...
Characterization of plant features is essential for effective management of plant growth monitoring, and precise management practices in crop production. Stereo vision captures multiple perspective images to create a threedimensional (3D) representation, enabling thorough analysis of crop structure and morphology. The objective of this study was to review the application of stereo vision in feature characterization of plants and fruit trees. Various features of plants such as height, canopy volume, plant spacing, intra-row spacing, and leaf area were surveyed for their characterization potential along with several data acquisition and data processing algorithms consisting of image segmentation, 3D image reconstruction, depth mapping, and disparity mapping. The study found out some results regarding the feature characterization of plants and fruit trees using stereovision. The tree canopy estimation results showed 6-7% error for elliptical and 2-3% error for conical shaped trees as well as for corn plants detection with an accuracy of 96.7% under natural light conditions. From a maximum distance of 5 cm and 1 cm, the errors were observed with the detection accuracy of 74.6% and 62.3%, respectively. The plant height of cabbage, potato, sesame, radish, and soybean were estimated with a R2 value of 0.78 to 0.84 and with an error less than 5%. Stereo vision achieved 97% precision with an RMSE of 0.016 m in wheat height measurement and distance of 20 m with errors below 5% for hazelnut trees. By addressing challenges and exploring various techniques, the paper concluded by summarizing key findings and suggesting directions for further research in plant growth and crop production practices.