Ciprian Pocol

3D lane detection system based on stereovision

This work presents a 3D lane detection method based on stereovision. The stereovision algorithm allows the elimination of the common assumptions: flat road, constant pitch angle or absence of roll angle. Moreover, the availability of 3D information allows the separation between the road and the obstacle features. The lane is modeled as a 3D surface, defined by the vertical and horizontal clothoid curves, the lane width and the roll angle. The lane detection is integrated into a tracking process. The current lane parameters are predicted using the past parameters and the vehicle dynamics, and this prediction provides search regions for the current detection. The detection starts with estimation of the vertical profile, using the stereo-provided 3D information, and afterwards the horizontal profile is detected using a model-matching technique in the image space, using the knowledge of the already detected vertical profile. The roll angle is detected last, by estimating the difference of the average heights of the left and right lane borders. The detection results are used to update the lane state through Kalman filtering.

High accuracy stereo vision system for far distance obstacle detection

This paper presents a high accuracy stereo vision system for obstacle detection and vehicle environment perception in a large variety of traffic scenarios, from highway to urban. The system detects obstacles of all types, even at high distance, outputting them as a list of cuboids having a position in 3D coordinates, size and speed.

A Sensor for Urban Driving Assistance Systems Based on Dense Stereovision

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.

Driving environment perception using stereovision

This paper presents a high accuracy, far range stereovision approach for driving environment perception based on 3D lane and obstacle detection. Stereovision allows the elimination of the common assumptions used in most monocular systems: flat road, constant pitch angle or absence of roll angle. The lane detection method is based on clothoidal 3D lane model. The detected lane parameters are the vertical and horizontal curvatures, the lane width and the roll angle. The detected lane profile is used for road obstacle features separation. Based on a vicinity criteria the over road 3D points are grouped and tracked over frames. The system detects and classifies the meaningful obstacles in terms of 3D position, size and speed.

Stereovision-Based Sensor for Intersection Assistance

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.

Spatial grouping of 3D points from multiple stereovision sensors

This paper presents a method for grouping 3D points into cuboids. The 3D points are extracted using multiple stereovision sensors, and the sensor fusion module performs the fusion of the data sets and the grouping of the points in a single algorithm. The fusion/grouping algorithm is scalable, being able to work using any number of sensors, including a single one. The grouping method relies on a method of transforming the 3D space so that the density of the points is kept constant, and all the points belonging to a single object are adjacent, making the grouping of points into cuboids a simple labeling problem.

Real-Time 3D Environment Reconstruction Using High Precision Trinocular Stereovision

This paper presents an implementation of a 3D environment reconstruction system, using trinocular (3 camera) stereovision. The system does not use rectification, in order to improve precision. Sub-pixel accuracy correlation is used. Feature extraction and correlation use MMX and SSE2 optimizations. Reconstruction correctitude tests were conducted using both synthetically generated images and camera acquired images

Shape improvement of traffic pedestrian hypotheses by means of stereo-vision and superpixels

Shape is a powerful descriptor frequently used in pedestrian detection process. This paper presents a novel stereo and superpixel-based approach for extracting high quality shapes of pedestrian hypotheses from urban traffic scenarios. Gray-levels stereo-vision images of traffic scenes are acquired, high quality stereo-reconstruction and optical flow algorithms are used for computing the depth and motion information. Superpixels are extracted using the intensity images and clustered in different obstacles by a novel paradigm that fuses intensity, depth and motion information. Pedestrian hypotheses are defined as a subset of the scene obstacles obtained by imposing human-specific geometric constraints. A contour tracing algorithm is used for extracting a continuous contour that defines the shape of each pedestrian hypothesis. A comparison between the contours quality of pedestrian hypotheses obtained by this stereo and superpixel approach and another approach based only on stereo-reconstructed points grouping shows improvements in both object shape description and area coverage. Improvements in shape description will increase the accuracy of any further pedestrian detection processes that use pattern matching techniques.