Junfeng Ge

Real-Time Pedestrian Detection and Tracking at Nighttime for Driver-Assistance Systems

Pedestrian detection is one of the most important components in driver-assistance systems. In this paper, we propose a monocular vision system for real-time pedestrian detection and tracking during nighttime driving with a near-infrared (NIR) camera. Three modules (region-of-interest (ROI) generation, object classification, and tracking) are integrated in a cascade, and each utilizes complementary visual features to distinguish the objects from the cluttered background in the range of 20-80 m. Based on the common fact that the objects appear brighter than the nearby background in nighttime NIR images, efficient ROI generation is done based on the dual-threshold segmentation algorithm. As there is large intraclass variability in the pedestrian class, a tree-structured, two-stage detector is proposed to tackle the problem through training separate classifiers on disjoint subsets of different image sizes and arranging the classifiers based on Haar-like and histogram-of-oriented-gradients (HOG) features in a coarse-to-fine manner. To suppress the false alarms and fill the detection gaps, template-matching-based tracking is adopted, and multiframe validation is used to obtain the final results. Results from extensive tests on both urban and suburban videos indicate that the algorithm can produce a detection rate of more than 90% at the cost of about 10 false alarms/h and perform as fast as the frame rate (30 frames/s) on a Pentium IV 3.0-GHz personal computer, which also demonstrates that the proposed system is feasible for practical applications and enjoys the advantage of low implementation cost.

Nighttime Pedestrian Detection with Near Infrared using Cascaded Classifiers

This paper presents a novel nighttime pedestrian detection approach only using a near infrared camera, which can be used in a practical driver assistance systems. This method can be divided into three steps: selection step, preprocess step and recognition step. Firstly, objects in the video are separated with an adaptive dual thresholds segmentation method in the selection step; Secondly, most of non-pedestrians are discarded with some constraints in the preprocess step; Finally, in the recognition step a cascaded classifiers with Histograms of Oriented Gradients and Adaptive Boosting Algorithm are introduced. Experiments on video sequences show that the proposed pedestrian detection approach has a high detection rate as well as a very low false alarm rate and run in real-time.

Novel light-leaking optical fiber liquid-level sensor for aircraft fuel gauging

Abstract. We present a novel light leaking optical fiber liquid level sensor for aircraft fuel measuring systems that is based on the side-emitting optical fiber. In operation, the fiber is bent as a spiral and inserted into the liquid, then the conducting light in the fiber will be scattered and transformed to be cladding mode and radiation mode. The radiation mode power in the scattering light is determined by the ambient refractive index. Since more power is lost in liquids than in the air, the output power of the sensor will monotonically decrease while the liquid level increases. The fiber is bent as a spiral to increase the sensor’s sensitivity. A laboratory prototype with a range of 1 m, a radius of 14 mm, and a screw pitch of 15 mm has been built and successfully demonstrated in water level measurement. The sensor’s response to liquid level is presented and shown to be consistent with the theory.

Continuous fuel level sensor based on spiral side-emitting optical fiber

A continuous fuel level sensor using a side-emitting optical fiber is introduced in this paper. This sensor operates on themodulation of the light intensity in fiber, which is caused by the claddings acceptance angle change when it is immersed in fuel. The fiber is bent as a spiral shape to increase the sensors sensitivity by increasing the attenuation coefficient and fibers submerged length compared to liquid level. The attenuation coefficients of fiber with different bent radiuses in the air and water are acquired through experiments. The fiber is designed as a spiral shape with a steadily changing slope, and its response to water level is simulated. The experimental results taken in water and aviation kerosene demonstrate a performance of 0.9m range and 10mm resolution.

Ice type detection using an oblique end-face fibre-optic technique

This paper reports on a novel, compound fibre-optic ice sensor consisting of two source fibre bundles and two signal ones. All the bundles had oblique fibre end-faces that were not perpendicular to the fibre axis. Dynamically icing experiments in an icing wind tunnel showed that with increasing thickness of ice accreted on the sensor, one group, comprising the first source and signal fibre bundles, output decreasing signal for glazed ice, increasing signal for rime ice, and constant signal for mixed ice, while the other group, comprising the second source and signal bundles, output constant signal for rime ice and decreasing signal for glazed or mixed ice. The different ice-growth curves are caused by different optical intensity distributions in the fibres for different ice types and thus can be used to accurately measure the ice type and thickness. This sensor is competitive in ice detection, especially identification of ice type in the aviation industry.

An optical fiber point liquid level sensor

An optical fiber sensor to accurately judge the level of a point is presented in this paper. Three fibers are used in the sensor, one as the source to emit infrared light, and the others as the differential inputs to absorb the scattering of light. Adopting a differential structure of dual signal fibers helps removing the interference of external environment and the change of the intensity of the light source. When the sensor contacts the liquid, the emitted light will be scatted by the liquid so that the output voltage will have significant difference from it when the sensor is above the liquid. A 45 degree oblique end-face of fibers is not just used to reduce the adhesion of liquid but also can increase discrepancy between differential inputs due to different scattered light intensity at different position, and then enhance the sensitivity. Base on this point, the optical fiber point liquid level sensor is designed and its working performance is investigated by experiments. The experimental results indicated stability, accuracy and excellent performance in the use of this sensor to measure the status of liquid level. And the safety will be ensured for the characteristic of fiber and excellent electrical isolation.