Osafumi Nakayama

Vehicle speed estimation using video data and acceleration information of a drive recorder

A drive recorder (DR) is installed in many vehicles to record information about car accidents. Recently, images and vehicle behavior data captured with DR is being used with increasing frequency to analyze dangerous driving behaviors and visualize the conditions in which this behavior occurs. Essential to this risk analysis is identifying the dangerous situation and obtaining information on the circumstances at the time, which must include precise vehicle speed information for analyzing detailed dangerous car motions such as rapid deceleration. Today, instead of high-precision speed pulse generator DR, many simple DR that record vehicle speeds captured by GPS are used, because they are easy to use without being connected to the car electronics. While these GPS-type DR are accurate enough for recording purposes, they are insufficient for analysis purposes, which hamper using DR for analysis in many cases. This study proposes a method that estimates precise vehicle speeds using images and acceleration data captured by simple DR. Most vehicle speed estimation methods calculate the velocity by estimating the vehicle motion that is represented by the motion flow of feature points in images. Maintaining the estimation accuracy requires obtaining these flow data from entire images. However, images captured by DR include few feature points of road surfaces that occupy the lower half of images and provide skewed feature-point flow position distributions, leading to a decrease in the accuracy of the speed estimation. The proposed method uses acceleration information, which is unique to DR, to calculate the area in which the vehicle can possibly move around, and restrict the image flow analysis to within this area. This can overcome skewed flow distributions to estimate the vehicle speed accurately. It was verified in actual running experiments that the proposed method significantly improves the accuracy of vehicle speed estimation compared with previous methods.

Robust Small-Object Detection for Outdoor Wide-Area Surveillance

In this paper, we present a robust small-object detection method, which we call “Frequency Pattern Emphasis Subtraction (FPES)”, for wide-area surveillance such as that of harbors, rivers, and plant premises. For achieving robust detection under changes in environmental conditions, such as illuminance level, weather, and camera vibration, our method distinguishes target objects from background and noise based on the differences in frequency components between them. The evaluation results demonstrate that our method detected more than 95% of target objects in the images of large surveillance areas ranging from 30–75 meters at their center.