Real-time unstructured road detection based on CNN and Gibbs energy function

Road detection algorithms with high robustness as well as timeliness are the basis for developing intelligent assisted driving systems. To improve the robustness as well as the timeliness of unstructured road detection, a new algorithm is proposed in this paper. First, for the first frame in the video, the homography matrix H is estimated based on the improved random sample consensus (RANSAC) algorithm for different regions in the image, and the features of H are automatically extracted using convolutional neural network (CNN), which in turn enables road detection. Secondly, in order to improve the rate of subsequent similar frame detection, the color as well as texture features of the road are extracted from the detection results of the first frame, and the corresponding Gaussian mixture models (GMMs) are constructed based on Orchard-Bouman, and then the Gibbs energy function is used to achieve road detection in subsequent frames. Finally, the above algorithm is verified in a real unstructured road scene, and the experimental results show that the algorithm is 98.4% accurate and can process 58 frames per second with 1024×960 pixels.

Automated near-field deformation detection from mobile laser scanning for the 2014 Mw 6.0 South Napa earthquake

Quantifying off-fault deformation in the near field remains a challenge for earthquake monitoring using geodetic observations. We propose an automated change detection strategy using geometric primitives generated using a deep neural network, random sample consensus and least squares adjustment. Using mobile laser scanning point clouds of vineyards acquired after the magnitude 6.0 2014 South Napa earthquake, our results reveal centimeter-level horizontal ground deformation over three kilometers along a segment of the West Napa Fault. A fault trace is detected from rows of vineyards modeled as planar primitives from the accumulated coseismic response, and the postseismic surface displacement field is revealed by tracking displacements of vineyard posts modeled as cylindrical primitives. Interpreted from the detected changes, we summarized distributions of deformation versus off-fault distances and found evidence of off-fault deformation. The proposed framework using geometric primitives is shown to be accurate and practical for detection of near-field off-fault deformation.

Surface Roughness Evaluation of the Inner Surface of Automobile Engine Bores by RANSAC and the Least Squares Method

Analysis methods for plateau surfaces have been described in the ISO standards, JIS, and previous studies. The authors of a previous study proposed a method based on the concept of random sample consensus (RANSAC). This method achieved high analysis accuracy for plateau surfaces by setting detailed conditions. However, the process of setting optimal conditions is performed manually, which reduces productivity due to the manpower and man-hours required. In this study, we propose a new method for automating the setting of conditions. This method, which does not require human intervention, is expected to contribute to the improvement of productivity at production sites.

A New and Direct R-Value Measurement Method of Sheet Metal Based on Multi-Camera DIC System

More and more attention has been given in the field of mechanical engineering to a material’s R-value, a parameter that characterizes the ability of sheet metal to resist thickness strain. Conventional methods used to determine R-value are based on experiments and an assumption of constant volume. Because the thickness strain cannot be directly measured, the R-value is currently determined using experimentally measured strains in the width, and loading directions in combination with the constant volume assumption, to determine the thickness strain indirectly. This paper provides an alternative method for determining the R-value without any assumptions. This method is based on the use of a multi-camera DIC system to measure strains in three directions simultaneously. Two sets of stereo-vision DIC measurement systems, each comprised of two GigE cameras, are placed on the front and back sides of the sample. Use of the double-sided calibration strategy unifies the world coordinate system of the front and back DIC measurement systems to one coordinate system, allowing for the measurement of thickness strain and direct calculation of R-value. The Random Sample Consensus (RANSAC) algorithm is used to eliminate noise in the thickness strain data, resulting in a more accurate R-value measurement.

Laser Scanning and Parametrization of Weld Grooves with Reflective Surfaces

This paper presents a novel weld groove parametrization algorithm, which is developed specifically for weld grooves in typical stub and butt joints between large tubular elements. The procedure is based on random sample consensus (RANSAC) with additionally proposed correction steps, including a corner correction step for grooves with narrow root weld, and an iterative error elimination step for improving the initially obtained data fit. The problem of curved groove sides (due to the pipe geometry) is attributed and solved. In addition, the procedure detects and eliminates several types of data noise due to laser line reflections. The performance of the procedure is studied experimentally using small-scale test objects, which have been ground using typical industrial power tools to achieve a realistic level of reflections. The execution times and data fit errors of the proposed procedure are compared to a procedure based on a more conventional RANSAC approach for line segment detection.

Research on Visual Positioning of a Roadheader and Construction of an Environment Map

The autonomous positioning of tunneling equipment is the key to intellectualization and robotization of a tunneling face. In this paper, a method based on simultaneous localization and mapping (SLAM) to estimate the body pose of a roadheader and build a navigation map of a roadway is presented. In terms of pose estimation, an RGB-D camera is used to collect images, and a pose calculation model of a roadheader is established based on random sample consensus (RANSAC) and iterative closest point (ICP); constructing a pose graph optimization model with closed-loop constraints. An iterative equation based on Levenberg–Marquadt is derived-, which can achieve the optimal estimation of the body pose. In terms of mapping, LiDAR is used to experimentally construct the grid map based on open-source algorithms, such as Gmapping, Cartographer, Karto, and Hector. A point cloud map, octree map, and compound map are experimentally constructed based on the open-source library RTAB-MAP. By setting parameters, such as the expansion radius of an obstacle and the updating frequency of the map, a cost map for the navigation of a roadheader is established. Combined with algorithms, such as Dijskra and timed-elastic-band, simulation experiments show that the combination of octree map and cost map can support global path planning and local obstacle avoidance.

Automatic Super-Surface Removal in Complex 3D Indoor Environments Using Iterative Region-Based RANSAC

Removing bounding surfaces such as walls, windows, curtains, and floor (i.e., super-surfaces) from a point cloud is a common task in a wide variety of computer vision applications (e.g., object recognition and human tracking). Popular plane segmentation methods such as Random Sample Consensus (RANSAC), are widely used to segment and remove surfaces from a point cloud. However, these estimators easily result in the incorrect association of foreground points to background bounding surfaces because of the stochasticity of randomly sampling, and the limited scene-specific knowledge used by these approaches. Additionally, identical approaches are generally used to detect bounding surfaces and surfaces that belong to foreground objects. Detecting and removing bounding surfaces in challenging (i.e., cluttered and dynamic) real-world scene can easily result in the erroneous removal of points belonging to desired foreground objects such as human bodies. To address these challenges, we introduce a novel super-surface removal technique for 3D complex indoor environments. Our method was developed to work with unorganized data captured from commercial depth sensors and supports varied sensor perspectives. We begin with preprocessing steps and dividing the input point cloud into four overlapped local regions. Then, we apply an iterative surface removal approach to all four regions to segment and remove the bounding surfaces. We evaluate the performance of our proposed method in terms of four conventional metrics: specificity, precision, recall, and F1 score, on three generated datasets representing different indoor environments. Our experimental results demonstrate that our proposed method is a robust super-surface removal and size reduction approach for complex 3D indoor environments while scoring the four evaluation metrics between 90% and 99%.