Automotive Research & Testing Center

Automotive Research & Test Center

Li-You Hsu

Hung-Pang Lin

Yu-Sheng Liao

Po-Hsiang Liao

Chih-Hung Yang

Ming-Fang Chang

Yu-Fang Wang

Tzu-Chien Hsu

Ying-Ren Chen

Yi-Shueh Tsai

Yu-Lung Chang

Siang-Min Siao

Hao-Han Chi

A lane stripe detecting method based on a three-dimensional LIDAR is for detecting a lane stripe of a road surface around a vehicle. A data acquisition transforming step is for obtaining a plurality of vehicle scan point coordinates. The vehicle scan point coordinates are divided into a plurality of scan lines, and each of the scan lines has a plurality of scan points. A horizontal layer lane stripe judging step is judging whether each of the scan points is a horizontal stripe point or a non-horizontal stripe point according to a threshold value of each of the scan lines. At least two of the threshold values of the scan lines are different from each other. A vertical layer lane stripe judging step is for judging whether each of the horizontal stripe points is a same lane stripe point or a different lane stripe point.

Lane stripe detecting method based on three-dimensional lidar and system thereof

A nighttime vehicle detecting method is for capturing an image by a camera and driving a computing unit to compute the image and then detect a highlight point of the image. The nighttime vehicle detecting method is for driving the computing unit to perform a communicating region labeling algorithm to label a plurality of highlight pixels connected to each other as a communicating region value, and then performing an area filtering algorithm to analyze an area of the highlight pixels connected to each other and judge whether the highlight pixels connected to each other are a vehicle lamp or not according to a size of the area. The nighttime vehicle detecting method is for driving the computing unit to perform an optical flow algorithm to obtain a speed of the vehicle lamp, and then filtering the vehicle lamp moved at the speed smaller than a predetermined speed.

Vehicle detecting method, nighttime vehicle detecting method based on dynamic light intensity and system thereof

The lidar detection device uses four two-dimensional lidars to scan an obstacle, to obtain original point-cloud data corresponding to the obstacle, and the original point-cloud data includes a relative distance, a relative angle and a relative speed of the obstacle relative to the vehicle, Next, the point-cloud data is classified into at least one point-cloud group corresponding to the obstacle, and a border length of the obstacle is obtained according to a contour of the point-cloud group. Kalman filter and extrapolation are used to estimate and track a movement path of a dynamic obstacle, and transmit the relative speed and the border length of the dynamic obstacle to an automatic driving controlling device. According to the relative distance, a coordinate of the dynamic obstacle nearest the vehicle can be obtained and transmitted to the automatic driving controlling device efficiently.

Lidar detection device of detecting close-distance obstacle and method thereof

A system for detecting a lane line of a road and a method thereof is disclosed. An image-retrieving device retrieves an image in front of a vehicle to generate a picture, and transmits the picture to a processing device. The processing device transforms one or two lane lines on the picture into a plurality of characteristic values, processes the characteristic values to generate one or two prediction lane lines, and uses them to dynamically predict and complete the lane lines, thereby forming a lane line prediction model. The processing device uses the lane line prediction model to determine whether the vehicle deviates, and transmits the picture and the lane line prediction model to a display device to display. When the vehicle deviates, the display device warns. The present invention can improve to recognize an unclear lane line, a single lane line and a far lane line.

System for detecting lane line of road and method thereof

A method for detecting driving behavior and a system using the same is disclosed. Firstly, at least one user image comprising an arm image and a head behavior image is retrieved. Then, a processor retrieves the arm image to compare with arm sample images stored in a database. The arm image corresponds to the arm sample image to determine a normal driving behavior. The arm image does not correspond to the arm sample image to determine an abnormal driving behavior, whereby the processor retrieves the head behavior image to compare with head regulation-violating sample images stored in the database. The head behavior image does not correspond to the head regulation-violating sample image to determine a normal driving behavior. The head behavior image corresponds to the head regulation-violating sample image to determine an abnormal driving behavior and output a second-level warning signal, thereby warning a driver.

Method for detecting driving behavior and system using the same

A dynamic lane line detection system and method is disclosed herein. Four wide-angle lenses are respectively arranged at four sides of a vehicle and capture distorted images. A vehicular image integration unit transforms the distorted images into an ordinary planar image and presents the planar image on an image display device together with an image of lane lines as information of lane departure warning for the driver. The system undertakes moving obstacle detection, around view monitor, or parking assistant according to images captured by four wide-angle lenses and the gearshift signal, speed signal, turning angle signal received by a vehicular signal receiving device. The present invention uses the wide-angle lenses and the vehicular image integration unit to effectively integrate lane departure information with a panoramic image, whereby to decrease the economical burden of the vehicle buyer and reduce the volume of hardware.

Dynamic lane line detection system and method

A self-adaptive image edge correction device and method thereof, including an image fetching unit, an image processing unit, and an image output unit. Wherein, the image fetching unit is used to provide an original image, and is connected electrically to the image processing unit, that includes a sharpening filter, a superimposer, and an edge detector. The sharpening filter converts the original image into a sharpened edge image, and the superimposer superimposes the original image to the sharpened edge image, to form an enhanced image. The edge detector fetches the edge of the enhanced image, to obtain a differential edge image. The image processing unit then utilizes selectively the horizontal correction or vertical correction to correct the differential edge image, based on s deviation direction of the differential edge image, to form a corrected image, and provide it to the image output unit to output as required.

Self-adaptive image edge correction device and method thereof

An automatic airview correction method comprises steps: moving a vehicle to an airview alignment pattern; capturing a plurality of airview alignment images of the surroundings of the vehicle; correcting distortion of the airview alignment images to obtain a plurality of corrected images; performing alignment compensation on the corrected images; searching for corner points of the corrected images and converting view points to obtain a plurality of view angle-converted images; and searching for corner points of the view angle-converted images, and seaming the view angle-converted images to form a panoramic airview and obtain parameters corresponding to the panoramic airview. The present invention can automatically align images and can also automatically detect corner points to seam the images of the surroundings of a vehicle, whereby to form a panoramic airview.

Automatic airview correction method

A self-adaptive image-based obstacle detection method comprises steps: capturing an original image; transforming the original image to an HSV color space, and retrieving a hue component (H) and a saturation component (S) of the HSV color space to form an HS-based image; dividing the HS-based image into image blocks; selecting one image block as a background block; using an obstacle recognition equation to determine whether each of the image blocks is similar to the background block; if no, deleting the image block; if yes, preserving the image block to form a binary obstacle image; and overlaying the binary obstacle image on the original image to filter out the background and obtain an initial ambit of an obstacle image. Then, three orderly movement flow equations are used to determine whether it is an obstacle.

Growing community of inventors

Changhua County, Taiwan

Li-You Hsu