Font Size: a A A

Design And Development Of Multi-source Image Data Fusion Platform

Posted on:2022-01-30Degree:MasterType:Thesis
Country:ChinaCandidate:Z ZhangFull Text:PDF
GTID:2518306572960709Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
With the development of the times and the advancement of human science and technology,different types of image sensors have emerged.The imaging mechanism of various image sensors,the range of operating wavelengths and the requirements for the working environment are different.Any image sensor with a single band is different.The collected images will have a certain degree of one-sidedness and cannot fully reflect the details and characteristics of the objects inside the image.Using multi-source sensors to collect image data for complementary information fusion can improve the utilization of image information.China's research on using multi-source sensors to obtain real target information under low light intensity mainly uses the fusion of medium and long wave infrared and visible light.Many studies in my country are only limited to the theoretical research stage,and there are few mature hardware experimental platforms.This paper designs a multi-source image data fusion platform based on the NVIDIA TX2 core processor,which meets the functions of synchronous triggering and acquisition of multi-channel high-definition cameras,large-scale data storage,complex algorithm operation,data transmission and connection with peripheral debugging equipment.At the same time,the research on the fusion of near-infrared and visible light images was carried out.According to the respective characteristics of near-infrared images and visible light images,image registration and fusion algorithms were studied.In the image registration,the feature points e xtracted by the ORB algorithm are screened twice.In order to meet the needs of more scenes,an improved multi-source image fusion algorithm based on the combination of image HIS color model transformation and wavelet transformation is proposed.Based on OpenCV The C++ program code builds a complete experimental system from near-infrared and visible light image acquisition to image registration and finally image fusion.The NVIDIA TX2 core processor core and device tree are re-modified,and the code is compiled and transplanted.Design and development of multi-source image data fusion platform based on TX2 core processor.
Keywords/Search Tags:Image fusion, Image registration, Near infrared image, Visible light image, TX2
PDF Full Text Request
Related items