Laksono Kurnianggoro

Traffic sign recognition system for autonomous vehicle using cascade SVM classifier

In past two decades, developing a system that can navigate vehicle autonomously becomes more interesting problem. The vehicle is equipped by sensors, such as radar, laser, GPS, and camera for sensing the surrounding. Among them, utilization of camera with computer vision technique is the most adopted method for constructing such a system. It is because camera provides a lot of information and is low-cost device rather than other sensors. Traffic road sign, as one of the important information from camera, carries a lot of useful information that are required for navigating. Thus, in this work, traffic sign detection and recognition is addressed. First, the input image is converted into normalize red and blue color space, as traffic sign usually appear with red and blue color. Second, maximally extremal stable region is then performed for extracting candidate region. Using heuristic rule of geometry properties, the false region will be excluded. Third, histogram of oriented gradient method is applied in order to extract feature from candidate region. Lastly, cascade support vector machine classifier is then processed to classify region belong to certain class of traffic sign. The extensive experiment would be carried out over German traffic sign recognition database and video. The experimental results demonstrate the effectiveness of our systems.

Camera and laser range finder fusion for real-time car detection

This paper describes a car detection method by combining data obtained from a laser and a camera. Data from the camera and the laser range finder (LRF) are combined after a calibration method has been performed. The calibration method defines the relative pose between camera and LRF. Car candidates are then extracted from the LRF data. The car candidate regions on the image are generated based on the filtered LRF data based on its size. To filter out the bad candidates, a verification method is performed on the car candidate regions. This method eliminates the needs of checking over several positions and scales, enables a speed enhancement over the general object detection strategy.

Scalable histogram of oriented gradients for multi-size car detection

This paper addresses two contributions for improving the accuracy and speed of preceding car detection systems. First, it proposes a feature description using Scalable Histogram of Oriented Gradient (SHOG) to solve scale problem of car region on the image. Without resizing the images to a fixed size, it is capable to extract a high-discriminated features with on the same feature space. Second, instead of use sliding window method to obtain candidate regions, it uses laser data information. This mechanism reduce the processing time significantly. In addition, an integral image method is utilized to support fast computation of the feature extraction. For classifying candidate regions into car and non-car class, linear support vector machine (SVM) is performed. The experimental results show that proposed descriptor accuracy is 3% higher than using standard HOG feature.

Real-Time Lane Region Detection Using a Combination of Geometrical and Image Features

Over the past few decades, pavement markings have played a key role in intelligent vehicle applications such as guidance, navigation, and control. However, there are still serious issues facing the problem of lane marking detection. For example, problems include excessive processing time and false detection due to similarities in color and edges between traffic signs (channeling lines, stop lines, crosswalk, arrows, etc.). This paper proposes a strategy to extract the lane marking information taking into consideration its features such as color, edge, and width, as well as the vehicle speed. Firstly, defining the region of interest is a critical task to achieve real-time performance. In this sense, the region of interest is dependent on vehicle speed. Secondly, the lane markings are detected by using a hybrid color-edge feature method along with a probabilistic method, based on distance-color dependence and a hierarchical fitting model. Thirdly, the following lane marking information is extracted: the number of lane markings to both sides of the vehicle, the respective fitting model, and the centroid information of the lane. Using these parameters, the region is computed by using a road geometric model. To evaluate the proposed method, a set of consecutive frames was used in order to validate the performance.

Ego-Motion Compensated for Moving Object Detection in a Mobile Robot

This paper presents a moving object detection method using optical flow in an image obtained from an omnidirectional camera mounted in a mobile robot. The moving object is extracted from the relative motion by segmenting the region representing the same optical flows after compensating the ego-motion of the camera. To obtain the optical flow, image is divided into grid windows and affine transformation is performed according to each window so that conformed optical flows are extracted. Moving objects are detected as transformed objects are different from the previously registered background. In omnidirectional and panoramic images, the optical flow seems to be emerging on focus of expansion FOE, on the contrary, it to be vanishing on focus of contraction FOC. FOE and FOC vectors are defined from the estimated optical flow and used as reference vectors for the relative evaluation of optical flow. In order to localize the moving objects, histogram vertical projection is applied with specific threshold. The algorithm was tested in a mobile robot and the proposed method achieved comparable results with 92.37% in detection rate.

Online Background-Subtraction with Motion Compensation for Freely Moving Camera

This paper proposes a background subtraction method for moving camera. The method relies on motion compensation to transfers the background model from the previous frame to the current frame. This motion compensation is carried out using homography transformation where the homography matrix is estimated from the set of point correspondences between previous and current frame. In order to achieve a fast processing speed, optical-flows from grid-based key-points are calculated to define the point correspondences. The background segmentation itself consists of 3 components: background model, candidate background model, and candidate age. Those 3 parameters are used to define the stable pixels which are considered as the background pixels. The proposed method was tested on a public benchmark system and achieved promising result as shown in the experimental report. Moreover, the method is able to work on real time with 56 fps of processing speed.

A Similarity-Based Approach for Shape Classification Using Region Decomposition

Measuring the similarity of two shapes is an important task in human vision systems in order to either recognize or classify the objects. For obtaining reliable results, a high discriminative shape descriptor should be extracted by considering both global and local information of the shape. Taking into account, this work introduces a centroid-based tree-structured (CENTREES) shape descriptor invariant to rotation and scale. Extracting the CENTREES descriptor is started by computing the central of mass of a binary shape, assigned as the root node of tree. The entire shape is then decomposed into b sub-shapes by voting each pixel point according to an angle between point and major principal axis relative to a centroid. In the same way, the central of mass of the sub-shapes are calculated and these locations are considered as level-1 nodes. These processes are repeated for a predetermined number of levels. For each node corresponding to sub-shapes, parameters invariant to translation, rotation and scale are extracted. A vector of all parameters is considered as descriptor. A feature-based template matching with X 2 distance function is used to measure shape dissimilarity. The evaluation of our descriptor is conducted using MPEG-7 dataset. The results justify that the CENTREES is one of reliable shape descriptors for shape similarity.

Estimation of collision risk for improving driver's safety

This paper introduces a method for analyzing the critical situation based on collision risk probability. Pedestrians in the scene are captured from a monocular camera mounted on the vehicle. Position information of object is extracted by projecting the centroid of bounding box to the ground plane. Five elements of collision criteria are used for our risk analysis. Pedestrian walking direction, its velocity and how aware pedestrian to the traffic are obtained from the pedestrian side. Car speed and relative distance of pedestrian from the car are extracted from car side. Then, with certain values of collision criteria, those elements are constructed. The critical situation is defined as joint probability of elements. Pedestrian are localized according to the critical situation as green for secure label, yellow for carefully and red for high priority to be alerted. A quantitative analysis is performed by measuring effectiveness of this approach. A real-world measurement and human perception survey are performed for evaluation. The performance evaluation shows our proposed method achieved average accuracy 87.5% and it significantly outperforms human perception survey with more than 30% improvement.

Iterative road detection based on vehicle speed

When moving towards fully autonomous navigation, safety plays the most important role for both pedestrian and driver. This paper proposes a method to estimate the lane road region of interest based on the stopping typical distance of a vehicle required by the current speed of the vehicle. This was achieved by taking advantage of the difference in color of the road surface given by the lane marking as well as the pavement road. This method was executed in three main steps. Firstly, a distance estimation method using the vehicle speed was presented. Secondly, a lane marking edge feature extraction method was proposed. Finally, in order to determine the road surface a curve fitting model was implemented. Preliminary results were performed and tested on a group of consecutive frames to prove the effectiveness of the proposed method.

Real-time lane marking detection

For autonomous navigation the real-time processing is crucial. This paper proposes a method to detect the lane markings in real-time using the advantage of parallel processing. A region of interest is constrained by the current velocity of a vehicle. The segmentation was achieved by utilizing a difference in color between lane marking and road pavement. The overall process is divided into three steps. The first is detection of lane markings based on the color probability. The second is the implementation of distance clustering analysis to define the surface course. Finally, The curve fitting was applied to assure the lane markings. The method was tested on a dataset to prove its effectiveness.