Font Size: a A A

Night Driver Assistance System Base On Infrared Thermal Imaging

Posted on:2014-06-09Degree:MasterType:Thesis
Country:ChinaCandidate:W Y LiangFull Text:PDF
GTID:2268330425975741Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
Infrared thermal imaging technology has been widely used in industrial, medical,defense and other fields with the developing of science and high technology inmaking infrared detector. This paper introduces a system of driving assistance in nightbase on infrared thermal imaging, its infrared detector that receives8~14μm infraredray with384*288resolution is made by ULIS company in France. The analog signalcoming from the infrared detector be sampled and changed to14bit digital signal by ahigh speed ADC. When a FPGA receives the infrared digital imaging signal willexecute Non-uniformity correction(NUC),dropt out bad pixels and enhance with theimage finally changes the image signal to BT656(4:2:2) and sends it to a videodecoder changed to CVBS displayed on a monitor.Human detected system is installed after infrared camera, it can be turned offwhen not used to reduce power. If the human detected is turned on, the CVBS signalfrom infrared camera will get into system and be changed to digital signal andprocessed by DSP, find out humans from the video and add warning message(ring uphuman with a rectangle),and finally the video be changed to CVBS again anddisplayed on LCD or other monitor. The video outputted from the DSP is real timeinfrared imaging with human detected ability. We must make sure that the thermalimaging is clear and be observed easily, with high contrast and SNR because humandetected is base on that, otherwise it would influence image segmentation and theeffect of human detected. The system developed an algorithm called digital detailenhancement (DDE) that improve effectively the image detail and the adaptive scene.Make full use of double threshold image segmentation, SVM classifier and modelmatch in human detected algorithm, application of multi-frame check effectivelyimprove the detection rate and reduce the false alarm rate. Experiments show thatthese algorithms have received very good results, the driver can clearly see the roadthrough the monitor, greatly improving driving safety!...
Keywords/Search Tags:infrared thermal imaging, image processing, human detected
PDF Full Text Request
Related items