Visible Image And Lidar 3D Research On Information Fusion Methods

Because the imaging laser radar can obtain the range image of the target and the accuracy is very high.Therefore,it is widely used in missile guidance,helicopter obstacle avoidance and 3D scene reconstruction.The effective fusion of visible light color image and lidar image can improve the overall performance of the system and provide more reliable images for subsequent visualization and target detection and recognition.For the realization of the visible light image and 3 d laser radar information fusion,based on the multi-resolution fusion algorithm as the core,through analyzing the characteristics of the laser radar image,to improve the rules of multi-resolution fusion algorithm to make it conform to the laser radar image fusion,and the simulation and experimental validation,through visible light image and 3 d laser radar information fusion,can obtain more comprehensive target information,and conforms to the human eye observation.The main research contents of this paper are as follows:First of all,through the analysis of imaging laser radar intensity like distance characteristics,determine the lidar intensity image and distance as first fusion,in order to obtain high quality intensity as the laser radar fusion,then the strength of the fusion image and visible light image fusion,so as to realize optical image and 3 d laser radar information fusion.Then,this paper studies the common image denoising algorithms,using the method based on off-line curvature estimation to complete the range image denoising,and then based on the SIFT feature extraction algorithm to register the visible image and the lidar intensity image,to complete the pre-processing before fusion.Second,through the simulation experiment analyzed the multi-resolution image fusion fusion algorithm in the process of laser radar image fusion existing problems,and then the fusion rules of the multi-resolution image fusion process was improved,both in the process of image restoration change intensity like distance like the top image weighted average fusion rule,directly using the top strength like as top level fusion result,make it more suitable for laser radar image fusion needs.Finally,the fusion experiment of visible light image and lidar image is carried out.The pixel-level information fusion experiments were carried out for indoor close-range scenes and outdoor long-distance scenes respectively,and the fusion results were evaluated by main evaluation and objective evaluation.The experimental fusion results show that the proposed fusion scheme and the improved multi-resolution fusion algorithm can improve the radar image quality while retaining the distance information and loading the color texture information.Adapted to lidar image fusion.