Semantic segmentation of target objects in power transmission line corridor point cloud scenes is a crucial step in powerline tree barrier detection. The massive quantity, disordered distribution, and non-uniformity of point clouds in power transmission line corridor scenes pose significant challenges for feature extraction. Previous studies have often overlooked the core utilization of spatial information, limiting the network\'s ability to understand complex geometric shapes. To overcome this limitation, this paper focuses on enhancing the deep expression of spatial geometric information in segmentation networks and proposes a method called BDF-Net to improve RandLA-Net. For each input 3D point cloud data, BDF-Net first encodes the relative coordinates and relative distance information into spatial geometric feature representations through the Spatial Information Encoding block to capture the local spatial structure of the point cloud data. Subsequently, the Bilinear Pooling block effectively combines the feature information of the point cloud with the spatial geometric representation by leveraging its bilinear interaction capability thus learning more discriminative local feature descriptors. The Global Feature Extraction block captures the global structure information in the point cloud data by using the ratio between the point position and the relative position, so as to enhance the semantic understanding ability of the network. In order to verify the performance of BDF-Net, this paper constructs a dataset, PPCD, for the point cloud scenario of transmission line corridors and conducts detailed experiments on it. The experimental results show that BDF-Net achieves significant performance improvements in various evaluation metrics, specifically achieving an OA of 97.16%, a mIoU of 77.48%, and a mAcc of 87.6%, which are 3.03%, 16.23%, and 18.44% higher than RandLA-Net, respectively. Moreover, comparisons with other state-of-the-art methods also verify the superiority of BDF-Net in point cloud semantic segmentation tasks.

This paper presents a novel segmentation algorithm specially developed for applications in 3D point clouds with high variability and noise, particularly suitable for heritage building 3D data. The method can be categorized within the segmentation procedures based on edge detection. In addition, it uses a graph-based topological structure generated from the supervoxelization of the 3D point clouds, which is used to make the closure of the edge points and to define the different segments. The algorithm provides a valuable tool for generating results that can be used in subsequent classification tasks and broader computer applications dealing with 3D point clouds. One of the characteristics of this segmentation method is that it is unsupervised, which makes it particularly advantageous for heritage applications where labelled data is scarce. It is also easily adaptable to different edge point detection and supervoxelization algorithms. Finally, the results show that the 3D data can be segmented into different architectural elements, which is important for further classification or recognition. Extensive testing on real data from historic buildings demonstrated the effectiveness of the method. The results show superior performance compared to three other segmentation methods, both globally and in the segmentation of planar and curved zones of historic buildings.

This paper proposes an autonomous robotic system to prune sweet pepper leaves using semantic segmentation with deep learning and an articulated manipulator. This system involves three main tasks: the perception of crop parts, the detection of pruning position, and the control of the articulated manipulator. A semantic segmentation neural network is employed to recognize the different parts of the sweet pepper plant, which is then used to create 3D point clouds for detecting the pruning position and the manipulator pose. Eventually, a manipulator robot is controlled to prune the crop part. This article provides a detailed description of the three tasks involved in building the sweet pepper pruning system and how to integrate them. In the experiments, we used a robot arm to manipulate the pruning leaf actions within a certain height range and a depth camera to obtain 3D point clouds. The control program was developed in different modules using various programming languages running on the ROS (Robot Operating System).

During recent years, hyperspectral imaging technologies have been widely applied in agriculture to evaluate complex plant physiological traits such as leaf moisture content, nutrient level, and disease stress. A critical component of this technique is white referencing used to remove the effect of non-uniform lighting intensity in different wavelengths on raw hyperspectral images. However, a flat white tile cannot accurately reflect the lighting intensity variance on plant leaves, since the leaf geometry (e.g., tilt angles) and its interaction with the illumination severely impact plant reflectance spectra and vegetation indices such as the normalized difference vegetation index (NDVI). In this research, the impacts of leaf angles on plant reflectance spectra were summarized, and an improved image calibration model using the fusion of leaf hyperspectral images and 3D point clouds was built. Corn and soybean leaf samples were imaged at different tilt angles and orientations using an indoor desktop hyperspectral imaging system and analyzed for differences in the NDVI values. The results showed that the leaf\'s NDVI largely changed with angles. The changing trends with angles differed between the two species. Using measurements of leaf tilt angle and orientation obtained from the 3D point cloud data taken simultaneously with the hyperspectral images, a support vector regression (SVR) model was successfully developed to calibrate the NDVI values of pixels at different angles on a leaf to a same standard as if the leaf was laid flat on a horizontal surface. The R-squared values between the measured and predicted leaf angle impacts were 0.76 and 0.94 for corn and soybean, respectively. This method has a potential to be used in any general plant imaging systems to improve the phenotyping quality.

This paper proposes a deep learning framework to encode subject-specific transformations between facial and bony shapes for orthognathic surgical planning. Our framework involves a bidirectional point-to-point convolutional network (P2P-Conv) to predict the transformations between facial and bony shapes. P2P-Conv is an extension of the state-of-the-art P2P-Net and leverages dynamic point-wise convolution (i.e., PointConv) to capture local-to-global spatial information. Data augmentation is carried out in the training of P2P-Conv with multiple point subsets from the facial and bony shapes. During inference, network outputs generated for multiple point subsets are combined into a dense transformation. Finally, non-rigid registration using the coherent point drift (CPD) algorithm is applied to generate surface meshes based on the predicted point sets. Experimental results on real-subject data demonstrate that our method substantially improves the prediction of facial and bony shapes over state-of-the-art methods.

We present 3DPointCaps++ for learning robust, flexible and generalizable 3D object representations without requiring heavy annotation efforts or supervision. Unlike conventional 3D generative models, our algorithm aims for building a structured latent space where certain factors of shape variations, such as object parts, can be disentangled into independent sub-spaces. Our novel decoder then acts on these individual latent sub-spaces (i.e. capsules) using deconvolution operators to reconstruct 3D points in a self-supervised manner. We further introduce a cluster loss ensuring that the points reconstructed by a single capsule remain local and do not spread across the object uncontrollably. These contributions allow our network to tackle the challenging tasks of part segmentation, part interpolation/replacement as well as correspondence estimation across rigid / non-rigid shape, and across / within category. Our extensive evaluations on ShapeNet objects and human scans demonstrate that our network can learn generic representations that are robust and useful in many applications.

During the last decades, consumer-grade RGB-D (red green blue-depth) cameras have gained popularity for several applications in agricultural environments. Interestingly, these cameras are used for spatial mapping that can serve for robot localization and navigation. Mapping the environment for targeted robotic applications in agricultural fields is a particularly challenging task, owing to the high spatial and temporal variability, the possible unfavorable light conditions, and the unpredictable nature of these environments. The aim of the present study was to investigate the use of RGB-D cameras and unmanned ground vehicle (UGV) for autonomously mapping the environment of commercial orchards as well as providing information about the tree height and canopy volume. The results from the ground-based mapping system were compared with the three-dimensional (3D) orthomosaics acquired by an unmanned aerial vehicle (UAV). Overall, both sensing methods led to similar height measurements, while the tree volume was more accurately calculated by RGB-D cameras, as the 3D point cloud captured by the ground system was far more detailed. Finally, fusion of the two datasets provided the most precise representation of the trees.

This study tested whether machine learning (ML) methods can effectively separate individual plants from complex 3D canopy laser scans as a prerequisite to analyzing particular plant features. For this, we scanned mung bean and chickpea crops with PlantEye (R) laser scanners. Firstly, we segmented the crop canopies from the background in 3D space using the Region Growing Segmentation algorithm. Then, Convolutional Neural Network (CNN) based ML algorithms were fine-tuned for plant counting. Application of the CNN-based (Convolutional Neural Network) processing architecture was possible only after we reduced the dimensionality of the data to 2D. This allowed for the identification of individual plants and their counting with an accuracy of 93.18% and 92.87% for mung bean and chickpea plants, respectively. These steps were connected to the phenotyping pipeline, which can now replace manual counting operations that are inefficient, costly, and error-prone. The use of CNN in this study was innovatively solved with dimensionality reduction, addition of height information as color, and consequent application of a 2D CNN-based approach. We found there to be a wide gap in the use of ML on 3D information. This gap will have to be addressed, especially for more complex plant feature extractions, which we intend to implement through further research.

In the current industrial landscape, increasingly pervaded by technological innovations, the adoption of optimized strategies for asset management is becoming a critical key success factor. Among the various strategies available, the \"Prognostics and Health Management\" strategy is able to support maintenance management decisions more accurately, through continuous monitoring of equipment health and \"Remaining Useful Life\" forecasting. In the present study, convolutional neural network-based deep neural network techniques are investigated for the remaining useful life prediction of a punch tool, whose degradation is caused by working surface deformations during the machining process. Surface deformation is determined using a 3D scanning sensor capable of returning point clouds with micrometric accuracy during the operation of the punching machine, avoiding both downtime and human intervention. The 3D point clouds thus obtained are transformed into bidimensional image-type maps, i.e., maps of depths and normal vectors, to fully exploit the potential of convolutional neural networks for extracting features. Such maps are then processed by comparing 15 genetically optimized architectures with the transfer learning of 19 pretrained models, using a classic machine learning approach, i.e., support vector regression, as a benchmark. The achieved results clearly show that, in this specific case, optimized architectures provide performance far superior (MAPE = 0.058) to that of transfer learning, which, instead, remains at a lower or slightly higher level (MAPE = 0.416) than support vector regression (MAPE = 0.857).

Leaf angle and leaf area index together influence canopy light interception and canopy photosynthesis. However, so far, there is no effective method to identify the optimal combination of these two parameters for canopy photosynthesis. In this study, first a robust high-throughput method for accurate segmentation of maize organs based on 3D point clouds data was developed, then the segmented plant organs were used to generate new 3D point clouds for the canopy of altered architectures. With this, we simulated the synergistic effect of leaf area and leaf angle on canopy photosynthesis. The results show that, compared to the traditional parameters describing the canopy photosynthesis including leaf area index, facet angle and canopy coverage, a new parameter - the canopy occupation volume (COV) - can better explain the variations of canopy photosynthetic capacity. Specifically, COV can explain > 79% variations of canopy photosynthesis generated by changing leaf angle and > 84% variations of canopy photosynthesis generated by changing leaf area. As COV can be calculated in a high-throughput manner based on the canopy point clouds, it can be used to evaluate canopy architecture in breeding and agronomic research.