Anselm Haselhoff

A vehicle detection system based on Haar and Triangle features

In recent years, the Viola and Jones rapid object detection approach became very popular. One aspect why this approach achieved acceptance is the numerical efficient computation of the Haar-like features on basis of the integral image. This efficiency is essential for sliding window techniques, where features have to be extracted for huge amounts of data. The main contribution of this paper is an efficient method to compute Triangle filters for feature extraction based on four integral images. The 2D Triangle filters are derived from 1D Bartlett functions. A comparison of Haar-like filters and the new Triangle filters is given by means of empirical results. The receiver operator characteristics reveal the superiority of the Triangle filters. Furthermore a vehicle detection system is described where the Triangle features are integrated. The system is based on a cascade of boosted classifiers, Haar and Triangle features, an adaptive sliding window and finally a Kalman filter.

A signal theoretic approach to measure the influence of image resolution for appearance-based vehicle detection

In this work a framework to measure the influence of training image resolution on classification performance for appearance-based object detection algorithms is presented. It is shown that based on sampling theory a reasonable image resolution for feature extraction can be chosen in advance, that is prior to the time consuming feature extraction and testing of the classifier. This is possible due to measuring the signal energy that is preserved in a low resolution image with respect to the optimal case of a high resolution image. The approach is justified using an AdaBoost algorithm with Haar-like features for vehicle detection. Tests of classifiers, trained with different resolutions, are performed and the results are presented. These results reveal that there is a good tradeoff between classification performance and computational load. The presented framework helps choosing a resolution for a good description of the training data.

An evolutionary optimized vehicle tracker in collaboration with a detection system

In this work a learning algorithm for visual object tracking is presented. As object representation a fast computable set of Haar-like features is used and a weighted correlation is applied for the matching process. The object tracker utilizes the same set of features that is already calculated for object detection and thus it is possible to reuse features for detection and tracking. The features weight values are optimized for the tracking purpose by means of evolutionary strategies. Different tests of the object tracker on real-world sequences are presented using vehicles as example objects. Additionally, an object detection system and the integration of the object tracker into that system is described. Besides the system is based on a cascade of boosted classifiers, Haar and Triangle features, an adaptive sliding window and finally a Kalman filter.

Radar-Vision Fusion with an Application to Car-Following using an Improved AdaBoost Detection Algorithm

This work outlines an object detection system for usage within driver assistance systems. The system detects vehicles that are driving ahead of an equipped vehicle by means of vision and radar data fusion. The radar provides a first estimation of vehicle candidates with related lateral position and distance information. This information is used to define a region of interest (ROI) on an image obtained by a video camera. An AdaBoost object detection algorithm is utilized to scan the ROI and verify radar detection. Due to the visual detection more specific data of the vehicles 3D position and width can be given. Moreover, the distance information provided by radar is used to choose optimal parameters during the visual detection process, e.g. properties of the scan window and parameters for fusing detections. In addition, this work will show that mutual information for haar-like feature selection can significantly increase detection rates using a new adaptive threshold.

Pedestrian crossing detecting as a part of an urban pedestrian safety system

Although recent statistics demonstrate a decrease in pedestrian fatalities, the absolute number of accident related deaths is sufficiently high to justify research in the area of vulnerable road user protection. Research has shown that situation awareness, which requires a significant amount of context information, is critical to timely intervention. Altered pedestrian behavior, due to traffic regulation, requires context information. For example, crosswalk presence and location knowledge can be of importance in pedestrian crossing scenarios. Therefore this paper discusses the implementation of a crosswalk detection algorithm, using Fourier transformation, augmented bipolarity, inverse perspective mapping template matching and edge orientation ratios for classification.

On visual crosswalk detection for driver assistance systems

In this paper a new crosswalk detection strategy is presented. The application area is narrowed down to driver assistance systems to guarantee a reliable detection result and to benefit from the properties of a vehicle mounted camera.

Multisensor data fusion for advanced driver assistance systems - the Active Safety Car project

Driver assistance systems support overstrained and affected drivers and become more and more essential for series-production vehicles. Object detection and segmentation is one of the most challenging research topics in this field. In order to warn the driver or automatically break before a potential collision, objects intersecting the path of the host vehicle have to be detected and classified. Most recently developed approaches are based on two dimensional image processing, sometimes in combination with a tracking algorithm associating detections in consecutive frames to one and the same object. Further robustness is achieved by multisensor data fusion, i.e. information by two or more different sensors (e.g. camera and radar data) are fused in order to get a much more reliable result. Another aspect for safety applications is communication between cars, which provides additional sensor locations and thus also requires data fusion technology. Two different approaches for data fusion are proposed and first results are presented.

2D line filters for vision-based lane detection and tracking

In this work a vision-based lane detection system is presented. The main contribution is the application of 2D line filters for lane detection which suppress noise in the near distance without destroying lines in the fare distance. Line pairs are decided to belong to lane markings by means of parallelism is world coordinates and reasonable constraints concerning the road width. The polygon constructed from the detected lane markings is then tracked over time via a kalman filter. Results of the approach are presented in terms of the detection accuracy on a labeled video sequence.

On Occlusion-Handling for People Detection Fusion in Multi-camera Networks

In this paper a system for people detection by means of Track-To-Track fusion of multiple cameras is presented. The main contribution of this paper is the evaluation of the fusion algorithm based on real image data. Before the fusion of the tracks an occlusion handling resolves implausible assignments.

On filtering by means of generalized integral images: a review and applications

In this paper the method of multidimensional (n-D) filtering based on prior signal integration is analyzed. This method has the advantage that the computational complexity for filtering is independent of the filter kernel size. An overview of recent 2-D image processing systems is presented where these types of filters are applied. Based on this overview a framework that covers this class of filters is derived using repeated integration. These filters include for example rect and triangle-filters which can be used to approximate Gaussian derivative filters. Furthermore the normalization of the filters, computational complexity, and storage cost are discussed. Finally, two image processing systems which benefit from the application of the filters are presented. They belong to the topic of advanced driver assistance systems.