Sergiu Nedevschi | Technical University of Cluj-Napoca (original) (raw)

Uploads

Papers by Sergiu Nedevschi

This paper presents an algorithm for tracking the cuboids generated from grouping the 3D points o... more This paper presents an algorithm for tracking the cuboids generated from grouping the 3D points obtained through stereovision. The solution described in the paper takes into consideration the particularities of the scenario and of the sensor, and brings considerable improvement in all the phases of tracking: initialization, prediction, measurement and update. The corner of the cuboid becomes the central working concept, thus improving the handling of partially occluded objects, of objects partially out of the field of view, and of objects whose measurement is fragmented by the sensor inaccuracies. After association at corner level, multiple measurements or validated parts of a measurement form a virtual object, the meta measurement, which is used for track update. The size of a vehicle is tracked using a histogram voting method. The resulted algorithm shows robustness and accuracy in the crowded urban scenario.

A new approach for the detection of the road surface and obstacles is presented. The 3D data from... more A new approach for the detection of the road surface and obstacles is presented. The 3D data from dense stereo is transformed into a rectangular elevation map. A quadratic road surface model is first fitted, by a RANSAC approach, to the region in front of the ego vehicle. This primary solution is then refined by a region growing-like process, driven by the 3D resolution and uncertainty model of the stereo sensor. An optimal global solution for the road surface is obtained. The road surface is used for a rough discrimination between road and above-road points. Above-road points are grouped based on vicinity and false areas are rejected. Each above-road area is classified into obstacles (cars, pedestrians etc.) or traffic isles (road-parallel patches) by using criteria related to the density of the 3D points. The proposed real-time algorithm was evaluated in an urban scenario and can be used in complex applications, from ego-pose estimation to path planning.

IEEE Transactions on Intelligent Transportation Systems, 2009

This paper presents a lane detection system that combines stereovision-specific techniques with g... more This paper presents a lane detection system that combines stereovision-specific techniques with grayscale image processing for maximizing the robustness and applicability against the difficult conditions of the urban environment. The lane marking features are extracted using a fast and robust dark-light-dark transition detector that's aware of the perspective effect. The clothoid lane model is matched to the extracted features using line segment fitting for two distance intervals, under special constraints that ensure correctness. Freeform lane border detection, independent on the geometry constraints, driven by lane marking features only, is used to solve the situations not suited for clothoid representation. The results of each detection method are fused together in a Kalman filter based framework.

IEEE Transactions on Intelligent Transportation Systems, 2009

The intersection scenario imposes radical changes in the physical setup and in the processing alg... more The intersection scenario imposes radical changes in the physical setup and in the processing algorithms of a stereo sensor. Due to the need for a wider field of view, which comes with distortions and reduced depth accuracy, increased accuracy in calibration and dense stereo reconstruction is required. The stereo matching process has to be performed on rectified images, by a dedicated stereo board, to free processor time for the high-level algorithms. In order to cope with the complex nature of the intersection, the proposed solution perceives the environment in two modes: a structured approach, for the scenarios where the road geometry is estimated from lane delimiters, and an unstructured approach, where the road geometry is estimated from elevation maps. The structured mode provides the parameters of the lane, and the position, size, speed and class of the static and dynamic objects, while the unstructured mode provides an occupancy grid having the cells labeled as free space, obstacle areas, curbs and isles.

This paper presents a lane estimation technique based on the particle filter framework, which fus... more This paper presents a lane estimation technique based on the particle filter framework, which fuses several image-based cues (edges, lane markings and curbs), and 3D cues extracted from stereovision. A partition sampling-like approach is used to decouple pitch estimation from the rest of the parameter set, allowing the use of a significantly lower number of particles, and initialization samples are used for faster handling of discontinuous roads. We also introduce a measure for detection quality, for result validation. The resulted solution has proven to be a reliable and fast lane detector for difficult scenarios.

IEEE Transactions on Vehicular Technology, 2010

The urban driving environment is a complex and demanding one, requiring increasingly complex sens... more The urban driving environment is a complex and demanding one, requiring increasingly complex sensors for the driving assistance systems. These sensors must be able to analyze the complex scene and extract all the relevant information, while keeping the response time as low as possible. The sensor presented in this paper answers to the requirements of the urban scenario through a multitude of detection modules, built on top of a hybrid (hardware plus software) dense stereo reconstruction engine. The sensor is able to detect and track clothoid and non-clothoid lanes, cars, pedestrians (classified as such), and drivable areas in the absence of lane markings. The hybrid stereovision engine and the proposed detection algorithms allow accurate sensing of the demanding urban scenario at a high frame rate.

This paper presents an algorithm for tracking the cuboids generated from grouping the 3D points o... more This paper presents an algorithm for tracking the cuboids generated from grouping the 3D points obtained through stereovision. The solution described in the paper takes into consideration the particularities of the scenario and of the sensor, and brings considerable improvement in all the phases of tracking: initialization, prediction, measurement and update. The corner of the cuboid becomes the central working concept, thus improving the handling of partially occluded objects, of objects partially out of the field of view, and of objects whose measurement is fragmented by the sensor inaccuracies. After association at corner level, multiple measurements or validated parts of a measurement form a virtual object, the meta measurement, which is used for track update. The size of a vehicle is tracked using a histogram voting method. The resulted algorithm shows robustness and accuracy in the crowded urban scenario.

A new approach for the detection of the road surface and obstacles is presented. The 3D data from... more A new approach for the detection of the road surface and obstacles is presented. The 3D data from dense stereo is transformed into a rectangular elevation map. A quadratic road surface model is first fitted, by a RANSAC approach, to the region in front of the ego vehicle. This primary solution is then refined by a region growing-like process, driven by the 3D resolution and uncertainty model of the stereo sensor. An optimal global solution for the road surface is obtained. The road surface is used for a rough discrimination between road and above-road points. Above-road points are grouped based on vicinity and false areas are rejected. Each above-road area is classified into obstacles (cars, pedestrians etc.) or traffic isles (road-parallel patches) by using criteria related to the density of the 3D points. The proposed real-time algorithm was evaluated in an urban scenario and can be used in complex applications, from ego-pose estimation to path planning.

IEEE Transactions on Intelligent Transportation Systems, 2009

This paper presents a lane detection system that combines stereovision-specific techniques with g... more This paper presents a lane detection system that combines stereovision-specific techniques with grayscale image processing for maximizing the robustness and applicability against the difficult conditions of the urban environment. The lane marking features are extracted using a fast and robust dark-light-dark transition detector that's aware of the perspective effect. The clothoid lane model is matched to the extracted features using line segment fitting for two distance intervals, under special constraints that ensure correctness. Freeform lane border detection, independent on the geometry constraints, driven by lane marking features only, is used to solve the situations not suited for clothoid representation. The results of each detection method are fused together in a Kalman filter based framework.

IEEE Transactions on Intelligent Transportation Systems, 2009

The intersection scenario imposes radical changes in the physical setup and in the processing alg... more The intersection scenario imposes radical changes in the physical setup and in the processing algorithms of a stereo sensor. Due to the need for a wider field of view, which comes with distortions and reduced depth accuracy, increased accuracy in calibration and dense stereo reconstruction is required. The stereo matching process has to be performed on rectified images, by a dedicated stereo board, to free processor time for the high-level algorithms. In order to cope with the complex nature of the intersection, the proposed solution perceives the environment in two modes: a structured approach, for the scenarios where the road geometry is estimated from lane delimiters, and an unstructured approach, where the road geometry is estimated from elevation maps. The structured mode provides the parameters of the lane, and the position, size, speed and class of the static and dynamic objects, while the unstructured mode provides an occupancy grid having the cells labeled as free space, obstacle areas, curbs and isles.

This paper presents a lane estimation technique based on the particle filter framework, which fus... more This paper presents a lane estimation technique based on the particle filter framework, which fuses several image-based cues (edges, lane markings and curbs), and 3D cues extracted from stereovision. A partition sampling-like approach is used to decouple pitch estimation from the rest of the parameter set, allowing the use of a significantly lower number of particles, and initialization samples are used for faster handling of discontinuous roads. We also introduce a measure for detection quality, for result validation. The resulted solution has proven to be a reliable and fast lane detector for difficult scenarios.

IEEE Transactions on Vehicular Technology, 2010

The urban driving environment is a complex and demanding one, requiring increasingly complex sens... more The urban driving environment is a complex and demanding one, requiring increasingly complex sensors for the driving assistance systems. These sensors must be able to analyze the complex scene and extract all the relevant information, while keeping the response time as low as possible. The sensor presented in this paper answers to the requirements of the urban scenario through a multitude of detection modules, built on top of a hybrid (hardware plus software) dense stereo reconstruction engine. The sensor is able to detect and track clothoid and non-clothoid lanes, cars, pedestrians (classified as such), and drivable areas in the absence of lane markings. The hybrid stereovision engine and the proposed detection algorithms allow accurate sensing of the demanding urban scenario at a high frame rate.