Research On Key Techniques Of Surface Bionic Compound Eye Velocity Measu Rement

The visual system is the most important functional organ for humans to perceive and understand the world.The information in the world includes both static and dynamic information,and humans can obtain most of the static and a small amount of dynamic information from the outside world through their visual organs.In nature,most organisms have different visual organs from humans,such as flies whose eyes are composed of multiple identical tiny lenses;The eyes of dragonflies are composed of tens of thousands of small eyes,and their structure belongs to a large biological compound eye.The eyes of organisms can not only obtain static information,but also a large amount of dynamic information.Therefore,biomimetic compound eyes have been widely used.The curved biomimetic compound eye imaging system has the characteristics of large field of view and high sensitivity to moving targets,making it a research hotspot for many domestic and foreign scholars.This article is based on a surface biomimetic compound eye camera developed in the laboratory.On the basis of large field imaging,a surface biomimetic compound eye imaging system is studied for measuring the motion speed of moving targets.Unlike ordinary single aperture imaging systems,the original image of the curved biomimetic compound eye is composed of 127 sub images.In order to conform to human visual habits,it is necessary to reconstruct 127 sub images into a complete image.Detect moving targets through relevant image processing algorithms and SIFT(Scale invariant feature transform)feature matching algorithms to calculate their motion speed.This article conducted experiments on close range low-speed targets,long-distance high-speed experiments,and comparative single aperture imaging systems.The specific research content includes:(1)Based on the velocity measurement principle of traditional single aperture imaging systems and the structural advantages of curved biomimetic compound eyes,a velocity measurement algorithm for curved biomimetic compound eye imaging systems is proposed.The original image of the compound eye is processed using MATLAB and a velocity measurement program is written.Firstly,the corresponding coordinate system was introduced.Starting from the optical principle of the curved surface biomimetic compound eye camera,a mathematical model of the imaging process was established,and the imaging process of the curved surface biomimetic compound eye was analyzed.Secondly,based on a good imaging foundation,the velocity measurement principle of the curved biomimetic compound eye imaging system is proposed.Finally,due to the presence of overlapping fields of view between adjacent sub eyes,a model of overlapping fields of view between adjacent sub eyes was established.By geometric transformation of the image,the computational burden of reducing the number of overlapping pixels is reduced,thereby calculating the motion speed of the moving object.(2)A novel image reconstruction algorithm based on Hough circle detection is proposed for curved biomimetic compound eye cameras.The traditional image stitching method is limited to stitching a small number of images.The original image of the compound eye is composed of 127 sub images.Each sub image is accurately located using the Hough circle detection algorithm,and combined with the inverse operation of the imaging process,127 sub images are reconstructed.In order to accurately calculate the number of overlapping pixels between adjacent sub images,the overlapping field of view of adjacent sub images is analyzed,and the SIFT feature point matching algorithm is used to extract the pixel coordinates of corresponding feature points.The overlapping field of view analysis mainly focuses on the horizontal adjacent two sub eyes,analyzing the relationship between object distance and the number of overlapping pixels.The SIFT feature point matching algorithm detects targets in the image,extracts the pixel coordinates of the target points,calculates the number of overlapping pixels,and provides more accurate information for later surface biomimetic compound eye velocity measurement.(3)Based on the principle of surface biomimetic compound eye velocity measurement,a velocity measurement experiment was designed.The experimental content is mainly divided into three parts,including: close range low-speed experiment with a compound eye camera,long-distance high-speed experiment with a compound eye camera,and speed measurement comparison experiment with a compound eye camera and a single aperture camera.The experimental results show that the traditional single aperture velocity measurement method has a velocity measurement error controlled within 8%.Compared with the single aperture velocity measurement method,the curved biomimetic compound eye velocity measurement method has a velocity measurement error controlled within 4%,resulting in higher velocity accuracy.Due to the advantage of overlapping fields of view between adjacent sub eyes in the curved biomimetic compound eye imaging system,the adjacent seven sub eyes can simultaneously obtain image information of the same target in the scene.Compared to traditional single aperture velocity measurement methods,the curved biomimetic compound eye velocity measurement method can obtain seven sets of measurement data in one experiment,effectively reducing the error caused by multiple experiments on the results.