DFSP: A fast and automatic distance field-based stem-leaf segmentation pipeline for point cloud of maize shoot

Abstract

The 3D point cloud data are used to analyze plant morphological structure. Organ segmentation of a single plant can be directly used to determine the accuracy and reliability of organ-level phenotypic estimation in a point-cloud study. However, it is difficult to achieve a high-precision, automatic, and fast plant point cloud segmentation. Besides, a few methods can easily integrate the global structural features and local morphological features of point clouds relatively at a reduced cost. In this paper, a distance field-based segmentation pipeline (DFSP) which could code the global spatial structure and local connection of a plant was developed to realize rapid organ location and segmentation. The terminal point clouds of different plant organs were first extracted via DFSP during the stem-leaf segmentation, followed by the identification of the low-end point cloud of maize stem based on the local geometric features. The regional growth was then combined to obtain a stem point cloud. Finally, the instance segmentation of the leaf point cloud was realized using DFSP. The segmentation method was tested on 420 maize and compared with the manually obtained ground truth. Notably, DFSP had an average processing time of 1.52 s for about 15,000 points of maize plant data. The mean precision, recall, and micro F1 score of the DFSP segmentation algorithm were 0.905, 0.899, and 0.902, respectively. These findings suggest that DFSP can accurately, rapidly, and automatically achieve maize stem-leaf segmentation tasks and could be effective in maize phenotype research. The source code can be found at https://github.com/syau-miao/DFSP.git.

Organ Segmentation and Phenotypic Trait Extraction of Cotton Seedling Point Clouds Based on a 3D Lightweight Network

“…It effectively extracts key phenotypic organs from point cloud data, and the segmentation results are closer to the actual shape. This study compares the segmentation algorithm based on PointNet++ and improved region growing (PRG-Net) used with other typical point cloud segmentation methods, including a DFSP segmentation [32], a density-based spatial clustering algorithm (DBSCAN) [33] and a clustering algorithm based on Euclidean distances (Euclidean Clustering) [34]. The specific comparison results, as shown in Table 5, indicate that across 10 test samples, the average precision of the algorithm in this paper reaches 96.67%, while the average precision of the other three algorithms are 75.4%, 70.33%, and 68.4%, respectively.…”

“…The neural network-based cotton seedling organ segmentation algorithm proposed in this paper accurately distinguishes the four leaves and the stem. In contrast, the DFSP segmentation [32] incorrectly segments the left leaf into three clusters; the density clustering algorithm [33] merges the top two leaves into one cluster; the Euclidean distance clustering algorithm [34] splits the largest leaf on the right into two clusters. Moreover, these three algorithms perform poorly in extracting stems because the branches and stems of the plant are closely connected, and traditional unsupervised learning methods struggle to effectively extract stem point cloud data.…”

“…Then, the neural network-based cotton seedling organ segmentation algorithm is used for stem and leaf point cloud segmentation. Compared to the average precision of 75.4% with the DFSP segmentation [32], the segmentation algorithm proposed in this paper achieves an average precision of 96.67%, realizing excellent segmentation of cotton seedlings. The calculated values of phenotypic parameters such as stem length, leaf length, leaf width, and leaf area have an average coefficient of determination R 2 of 92.71%, with the maximum root mean square error (RMSE) reaching only 4.6306 cm 2 and a mean percentage error (MAPE) of 0.1069%, ensuring the precision of the extracted phenotypic information from cotton seedlings.…”

“…Accurate segmentation of plant stems and leaves is crucial for precisely measuring phenotypic parameters associated with different organs [24]. A neural network-based segmentation algorithm for cotton seedling organs is proposed, which shows higher adapt-ability compared to the unsupervised DFSP segmentation algorithm [32]. It directly extracts features from raw data and completes the segmentation task through end-to-end learning.…”

Three-dimensional branch segmentation and phenotype extraction of maize tassel based on deep learning

“…The DBSCAN(density-based spatial clustering of applications with noise) algorithm is applied to automatic branch segmentation of maize tassel point cloud, but it was difficult to achieve branching and segmentation for compact tassels, so more robust algorithms need to be studied [ 22 ]. The DFSP(distance field-based segmentation pipeline) algorithm was proposed for automated segmentation of corn plant stem and leaf point clouds in different directional structures [ 23 ].…”