Font Size: a A A

A Research Of Vehicle Safety Based On Infrared Technology

Posted on:2017-05-29Degree:MasterType:Thesis
Country:ChinaCandidate:J FanFull Text:PDF
GTID:2272330485988289Subject:Optical Engineering
Abstract/Summary:PDF Full Text Request
Recently,the number of cars increased gradually with the increasing improvement of people’s living standards.The vehicle safety issues has been widespread concern.It is prone to accidents because of the poor lighting and the unclear road at night,So,the driving at night is a safety issue and it has been a research focus.With the development of uncooled infrared focal plane,all-weather observation using infrared imaging systems embedded in the car which can penetrates the darkness and smoke will help improve traffic safety of the vehicle.Therefore,vehicle safety technology based on infrared is of great practical significance.This paper take the infrared video sequences pedestrian as the research object.The purpose is to complete the study and simulation of pedestrian detection algorithm and completed the development of automotive systems primarily for collision warning.In the simulation study of pedestrian detection,this paper first do some introduction for the present situation of the pedestrian detection.And the paper elaborated the principles of the pedestrian detection technology and systems.Morever,various methods for image segmentation is been comparative analysis by the tool of Matlab and a self-thresholding clustering analysis which can compete the image segmentation has been put forward in this paper.Then,the paper analysis the various features of the infrared image and select the direction histogram feature infrared image sequences as the feature.Finally,pedestrian detection and identification has been completed via a model which produce from machine training and machine learning.In a study of automotive collision warning system,a system based on ARM has been designed.Functions such as collision warning navigation and communication has been developed based on the operating system of WINCE.The subject adopts Universal U-blox GPS modules NEO-6 Series to get vehicle position information and use gravity sensor MPU-6050 gyro sensors to obtain the vehicle attitude information.Informations transmit to the processor which complete GPS information decoding and attitude angle solver by way of the serial port.The paper also defines the criteria for judging the collision,complete vehicle collision detection and alarm.
Keywords/Search Tags:infrared image sequences, pedestrian detection, image segmentation, machine learning, vehicle warning systems
PDF Full Text Request
Related items