Automated Object Detection, Mapping, and Assessment of Roadside Clear Zones Using Lidar Data

This paper proposes a fully automated approach to map and assess roadside clearance parameters using mobile Light Detection and Ranging (lidar) data on rural highways. Compared with traditional manual surveying methods, lidar data could provide a more efficient and cost-effective source to extract roadside information. This study proposes a novel voxel-based raycasting approach focused primarily on automating roadside mapping and assessment. First, the scanning vehicle trajectory is extracted. Pavement surface points are then detected, and a method is proposed to extract pavement edge trajectories. Once pavement edges are extracted, guardrails were identified using a conical frustum emitted from the edge trajectory points. Target points and flexion points are then generated and located on the roadside, and a voxel-based raycasting approach is used to search for roadside obstacles and query their locations. Finally, roadside slopes and embankment heights were mapped at specific intervals, and roadside design guidelines and requirements were automatically checked against the mapping results. Noncompliant locations with substandard conditions were automatically queried. The method was tested on four highway segments in Alberta, Canada. The accuracy of the edge detection reached up to 98.5%. Furthermore, the method proved to be accurate in object detection, being able to detect all obstructions on the roadside in each tested segment. The proposed method can help transportation authorities automatically map and inventory roadside clearance parameters. Moreover, the safety performance of existing road infrastructure can be studied using collected information and crash data to support decision making on road maintenance and upgrades.

Segmentation of Drilled Holes in Texture Wooden Furniture Panels Using Deep Neural Network

Drilling operations are an essential part of furniture from MDF laminated boards required for product assembly. Faults in the process might introduce adverse effects to the furniture. Inspection of the drilling quality can be challenging due to a big variety of board surface textures, dust, or woodchips in the manufacturing process, milling cutouts, and other kinds of defects. Intelligent computer vision methods can be engaged for global contextual analysis with local information attention for automated object detection and segmentation. In this paper, we propose blind and through drilled holes segmentation on textured wooden furniture panel images using the UNet encoder-decoder modifications enhanced with residual connections, atrous spatial pyramid pooling, squeeze and excitation module, and CoordConv layers for better segmentation performance. We show that even a lightweight architecture is capable to perform on a range of complex textures and is able to distinguish the holes drilling operations’ semantical information from the rest of the furniture board and conveyor context. The proposed model configurations yield better results in more complex cases with a not significant or small bump in processing time. Experimental results demonstrate that our best-proposed solution achieves a Dice score of up to 97.89% compared to the baseline U-Net model’s Dice score of 94.50%. Statistical, visual, and computational properties of each convolutional neural network architecture are addressed.

An Advanced Approach to Detect Edges of Digital Images for Image Segmentation

Image segmentation has been an active topic of research for many years. Edges characterize boundaries, and therefore, detection of edges is a problem of fundamental importance in image processing. Edge detection in images significantly reduces the amount of data and filters out useless information while preserving the important structural properties in an image. Edges carry significant information about the image structure and shape, which is useful in various applications related with computer vision. In many applications, the edge detection is used as a pre-processing step. Edge detection is highly beneficial in automated cell counting, structural analysis of the image, automated object detection, shape analysis, optical character recognition, etc. Different filters are developed to find the gradients and detect edges. In this chapter, a new filter (kernel) is proposed, and the compass operator is applied on it to detect edges more efficiently. The results are compared with some of the previously proposed filters both qualitatively and quantitatively.