Abstract

The article considers the active safety system of road freight transport. The stereoscopic computer vision is the core of the system. The article also describes the major algorithms of active safety and the accuracy characteristics of algorithms’ application. Document type: Part of book or chapter of book

Abstract

This paper discusses the detection of moving objects (being a crucial part of driver assistance systems) using monocular or stereoscopic computer vision. In both cases, object detection is based on motion analysis of individually tracked image points (optical flow), providing a motion metric which corresponds to the likelihood that the tracked point is moving. Based on this metric, points are segmented into objects by employing a globally optimal graph-cut algorithm. Both approaches are comparatively evaluated using real-world vehicle image sequences. Document type: Part of book or chapter of book

Abstract

Automatic multi-class object detection in remote sensing images in unconstrained scenarios is of high interest for several applications including traffic monitoring and disaster management. The huge variation in object scale, orientation, category, and complex backgrounds, as well as the different camera sensors pose great challenges for current algorithms. In this work, we propose a new method consisting of a novel joint image cascade and feature pyramid network with multi-size convolution kernels to extract multi-scale strong and weak semantic features. These features are fed into rotation-based region proposal and region of interest networks to produce object detections. Finally, rotational non-maximum suppression is applied to remove redundant detections. During training, we minimize joint horizontal and oriented bounding box loss functions, as well as a novel loss that enforces oriented boxes to be rectangular. Our method achieves 68.16% mAP on horizontal and 72.45% mAP on oriented bounding box detection tasks on the challenging DOTA dataset, outperforming all published methods by a large margin (+6% and +12% absolute improvement, respectively). Furthermore, it generalizes to two other datasets, NWPU VHR-10 and UCAS-AOD, and achieves competitive results with the baselines even when trained on DOTA. Our method can be deployed in multi-class object detection applications, regardless of the image and object scales and orientations, making it a great choice for unconstrained aerial and satellite imagery. Document type: Part of book or chapter of book

Abstract

In order to make machines perceive their external environment coherently, multiple sources of sensory information derived from several different modalities can be used (e.g. cameras, LIDAR, stereo, RGB-D, and radars). All these different sources of information can be efficiently merged to form a robust perception of the environment. Some of the mechanisms that underlie this merging of the sensor information are highlighted in this chapter, showing that depending on the type of information, different combination and integration strategies can be used and that prior knowledge are often required for interpreting the sensory signals efficiently. The notion that perception involves Bayesian inference is an increasingly popular position taken by a considerable number of researchers. Bayesian models have provided insights into many perceptual phenomena, showing that they are a valid approach to deal with real-world uncertainties and for robust classification, including classification in time-dependent problems. This chapter addresses the use of Bayesian networks applied to sensory perception in the following areas: mobile robotics, autonomous driving systems, advanced driver assistance systems, sensor fusion for object detection, and EEG-based mental states classification. Document type: Part of book or chapter of book

Abstract

Regular inspection of underwater communications (pipelines and cables) is actual problem of modern oil and gas industry. Specially equipped vessels, towed underwater devices and remote operated vehicles /ROV/ are applied for these purposes as usually, but quality of acquired data does not always allow revealing emergencies at the proper time. “Spot” inspections by ROVs give difficultly comparable data (Baker, 1991; Murray, 1991). The perspective solution of the problem is autonomous underwater vehicles /AUV/ application as “the intellectual carrier” of research equipment (Evans et al., 2003; Kojima et al., 1997). According (Ageev, 2005) the main goals of pipeline and cables inspection are: 1. more accurate position determination (searching and tracking); 2. pipe sagging and freespan detection and measurement; 3. terrain survey on each side of communication by means of high frequency side scan sonar /HF SSS/ and detection of extraneous objects; 4. detection of damages; 5. leakage detection of transported substances (for pipelines). The pipeline and cable inspection by means of AUV includes two stages: preliminary (communication search and detection) and the main (motion along the communication with carrying out of necessary measurements, i.e. tracking). Exact mutual orientation of AUV and inspected object is required in real time during the tracking stage. To solve inspection tasks AUV should be equipped with reliable detection systems for inspected object recognition. Video, electromagnetic and echo-sounder data can be used for these purposes. Each of these devices demonstrates optimal results for certain classes of objects in appropriate conditions. For example, metal pipelines have the significant sizes and can be detected by all listed above devices. While underwater cables have a small diameter, because of this applicability of acoustic methods is limited (Petillot et al., 2002). Process of communications search and detection is complicated, as a rule, with a poor visibility of the given objects (strewed with a ground, silted or covered by underwater flora and fauna). Experiments with the use of AUV for inspection of underwater communications have been carried out for a long time. Usually only one instrument, which AUV is equipped with, is used for object detection. O pe n A cc es s D at ab as e w w w .in te ch w eb .o rg Document type: Part of book or chapter of book

Abstract

In this paper, we present a deep neural network based real-time integrated framework to detect objects, lane markings, and drivable space using a monocular camera for advanced driver assistance systems. The object detection framework detects and tracks objects on the road such as cars, trucks, pedestrians, bicycles, motorcycles, and traffic signs. The lane detection framework identifies the different lane markings on the road and also distinguishes between the ego lane and adjacent lane boundaries. The free space detection framework estimates the drivable space in front of the vehicle. In our integrated framework, we propose a pipeline combining the three deep neural networks into a single framework, for object detection, lane detection, and free space detection simultaneously. The integrated framework is implemented in C++ and runs real-time on the Nvidia's Drive PX 2 platform.

Abstract

utomation of disassembly processes in electronic waste recycling is progressing but hindered by the lack of automated procedures for screw detection and removal. Here we specifically address the detection problem and implement a universal, generalizable, and extendable screw detector which can be deployed in automated disassembly lines. We selected the best performing state-of-the-art classifiers and compared their performance to that of our architecture, which combines a Hough transform with a novel integrated model of two deep convolutional neural networks for screw detection. We show that our method outperforms currently existing methods, while maintaining the high speed of computation. Data set and code of this study are made public.

Abstract

n underwater imaging system was investigated for automotive use in highly scattered underwater environments. The purpose of the system is the driver&rsquo s information about hidden obstacles, such as stones, driftwood, open sewer hatches. A comparison of various underwater vision methods was presented by the way they are implemented, the range reached, and the cost of implementation. It has been experimentally shown that a conventional active system can provide a maximum visibility range of up to three light attenuation lengths. In most practical cases of turbid waters during floods, this corresponds to distances of about 1 meter. From the presented analysis it follows that advanced extended range imaging methods allow increasing of the visibility range up to 2 meters. Document type: Article