Search

A validation data set plays a pivotal role in tweaking a machine learning model trained in a supervised manner. Many existing algorithms select a part of available data by using random sampling to produce a validation set. However, this approach can be prone to overfitting. One should follow careful data splitting to have reliable training and validation sets that can produce a generalized model with a good performance for the unseen (test) data. Data splitting based on resampling techniques involves repeatedly drawing samples from the available data. Hence, resampling methods can give better generalization power to a model, because they can produce and use many training and/or validation sets. These techniques are computationally expensive, but with increasingly available high-performance computing facilities, one can exploit them. Though a multitude of resampling methods exist, investigation of their influence on the generality of deep learning (DL) algorithms is limited due to its non-linear black-box nature. This paper contributes by: (1) investigating the generalization capability of the four most popular resampling methods: k-fold cross-validation (k-CV), repeated k-CV (Rk-CV), Monte Carlo CV (MC-CV) and bootstrap for creating training and validation data sets used for developing, training and validating DL based point cloud classifiers (e.g., PointNet; Qi et al., 2017a), (2) justifying Mean Square Error (MSE) as a statistically consistent estimator, and (3) exploring the use of MSE as a reliable performance metric for supervised DL. Experiments in this paper are performed on both synthetic and real-world aerial laser scanning (ALS) point clouds. [ABSTRACT FROM AUTHOR]

Road surface extraction is crucial for 3D city analysis. Mobile laser scanning (MLS) is the most appropriate data acquisition system for the road environment because of its efficient vehicle-based on-road scanning opportunity. Many methods are available for road pavement, curb and roadside way extraction. Most of them use classical approaches that do not mitigate problems caused by the presence of noise and outliers. In practice, however, laser scanning point clouds are not free from noise and outliers, and it is apparent that the presence of a very small portion of outliers and noise can produce unreliable and non-robust results. A road surface usually consists of three key parts: road pavement, curb and roadside way. This paper investigates the problem of road surface extraction in the presence of noise and outliers, and proposes a robust algorithm for road pavement, curb, road divider/islands, and roadside way extraction using MLS point clouds. The proposed algorithm employs robust statistical approaches to remove the consequences of the presence of noise and outliers. It consists of five sequential steps for road ground and non-ground surface separation, and road related components determination. Demonstration on two different MLS data sets shows that the new algorithm is efficient for road surface extraction and for classifying road pavement, curb, road divider/island and roadside way. The success can be rated in one experiment in this paper, where we extract curb points; the results achieve 97.28%, 100% and 0.986 of precision, recall and Matthews correlation coefficient, respectively. [ABSTRACT FROM AUTHOR]

Most deep learning (DL) methods that are not end-to-end use several multi-scale and multi-type hand-crafted features that make the network challenging, more computationally intensive and vulnerable to overfitting. Furthermore, reliance on empirically-based feature dimensionality reduction may lead to misclassification. In contrast, efficient feature management can reduce storage and computational complexities, builds better classifiers, and improves overall performance. Principal Component Analysis (PCA) is a well-known dimension reduction technique that has been used for feature extraction. This paper presents a two-step PCA based feature extraction algorithm that employs a variant of feature-based PointNet (Qi et al., 2017a) for point cloud classification. This paper extends the PointNet framework for use on large-scale aerial LiDAR data, and contributes by (i) developing a new feature extraction algorithm, (ii) exploring the impact of dimensionality reduction in feature extraction, and (iii) introducing a non-end-to-end PointNet variant for per point classification in point clouds. This is demonstrated on aerial laser scanning (ALS) point clouds. The algorithm successfully reduces the dimension of the feature space without sacrificing performance, as benchmarked against the original PointNet algorithm. When tested on the well-known Vaihingen data set, the proposed algorithm achieves an Overall Accuracy (OA) of 74.64% by using 9 input vectors and 14 shape features, whereas with the same 9 input vectors and only 5PCs (principal components built by the 14 shape features) it actually achieves a higher OA of 75.36% which demonstrates the effect of efficient dimensionality reduction. [ABSTRACT FROM AUTHOR]

Semantic segmentation of point clouds is indispensable for 3D scene understanding. Point clouds have credibility for capturing geometry of objects including shape, size, and orientation. Deep learning (DL) has been recognized as the most successful approach for image semantic segmentation. Applied to point clouds, performance of the many DL algorithms degrades, because point clouds are often sparse and have irregular data format. As a result, point clouds are regularly first transformed into voxel grids or image collections. PointNet was the first promising algorithm that feeds point clouds directly into the DL architecture. Although PointNet achieved remarkable performance on indoor point clouds, its performance has not been extensively studied in large-scale outdoor point clouds. So far, we know, no study on large-scale aerial point clouds investigates the sensitivity of the hyper-parameters used in the PointNet. This paper evaluates PointNet's performance for semantic segmentation through three large-scale Airborne Laser Scanning (ALS) point clouds of urban environments. Reported results show that PointNet has potential in large-scale outdoor scene semantic segmentation. A remarkable limitation of PointNet is that it does not consider local structure induced by the metric space made by its local neighbors. Experiments exhibit PointNet is expressively sensitive to the hyper-parameters like batch-size, block partition and the number of points in a block. For an ALS dataset, we get significant difference between overall accuracies of 67.5% and 72.8%, for the block sizes of 5m × 5m and 10m × 10m, respectively. Results also discover that the performance of PointNet depends on the selection of input vectors. [ABSTRACT FROM AUTHOR]

Semantic segmentation of point clouds is indispensable for 3D scene understanding. Point clouds have credibility for capturing geometry of objects including shape, size, and orientation. Deep learning (DL) has been recognized as the most successful approach for image semantic segmentation. Applied to point clouds, performance of the many DL algorithms degrades, because point clouds are often sparse and have irregular data format. As a result, point clouds are regularly first transformed into voxel grids or image collections. PointNet was the first promising algorithm that feeds point clouds directly into the DL architecture. Although PointNet achieved remarkable performance on indoor point clouds, its performance has not been extensively studied in large-scale outdoor point clouds. So far, we know, no study on large-scale aerial point clouds investigates the sensitivity of the hyper-parameters used in the PointNet. This paper evaluates PointNet's performance for semantic segmentation through three large-scale Airborne Laser Scanning (ALS) point clouds of urban environments. Reported results show that PointNet has potential in large-scale outdoor scene semantic segmentation. A remarkable limitation of PointNet is that it does not consider local structure induced by the metric space made by its local neighbors. Experiments exhibit PointNet is expressively sensitive to the hyper-parameters like batch-size, block partition and the number of points in a block. For an ALS dataset, we get significant difference between overall accuracies of 67.5% and 72.8%, for the block sizes of 5m × 5m and 10m × 10m, respectively. Results also discover that the performance of PointNet depends on the selection of input vectors. [ABSTRACT FROM AUTHOR]

Ground surface extraction is one of the classic tasks in airborne laser scanning (ALS) point cloud processing that is used for three-dimensional (3D) city modelling, infrastructure health monitoring, and disaster management. Many methods have been developed over the last three decades. Recently, Deep Learning (DL) has become the most dominant technique for 3D point cloud classification. DL methods used for classification can be categorized into end-to-end and non end-to-end approaches. One of the main challenges of using supervised DL approaches is getting a sufficient amount of training data. The main advantage of using a supervised non end-to-end approach is that it requires less training data. This paper introduces a novel local feature-based non end-to-end DL algorithm that generates a binary classifier for ground point filtering. It studies feature relevance, and investigates three models that are different combinations of features. This method is free from the limitations of point clouds' irregular data structure and varying data density, which is the biggest challenge for using the elegant convolutional neural network. The new algorithm does not require transforming data into regular 3D voxel grids or any rasterization. The performance of the new method has been demonstrated through two ALS datasets covering urban environments. The method successfully labels ground and non-ground points in the presence of steep slopes and height discontinuity in the terrain. Experiments in this paper show that the algorithm achieves around 97% in both F1-score and model accuracy for ground point labelling. [ABSTRACT FROM AUTHOR]