Research On Key Technologies Of Multi-channel Image Fusion

At present,traditional industrial vision products can only obtain limited information under a single spectrum in the process of observing scenes and targets due to their shortcomings of single imaging spectrum.Also,they can not achieve extensive coverage of multi-dimensional target information in scenes.For example,this process is siginifanctly affected due to its poor performance in the relatively dark environment.It should be noted that it is impossible to quickly distinguish the target morphology of abnormal heating if we only use infrared imaging to capture the characteristics of the object.Besides,the characteristic of ultraviolet light imaging in the traditional high voltage arc detection lead to the fact that it is difficult to locate the position of the high voltage power equipment that sends the arc quickly.At present,it is highly important to adapt the current techniques.But most researchers have not delved deep enough.To resolve the issue that a variety of spectral equipment must be carried,which is not only cumbersome to operate,but also inconvenient to carry,but also increases the use cost and brings a lot of inconvenience,the emergence of innovative equipment that can integrate multiple spectral features and distinguish them in the primary imaging process has become a top priority in the field of imaging vision.A multi-spectral fusion intelligent imaging system is designed and implemented to meet the urgent need.This system integrates three aspects of optical components,investigating specific algorithms and corresponding system design.Through the pixel-level fusion algorithm,it can easily collect information from spectrum that spans the 10400 nm band of visible light,infrared light and ultraviolet light.The whole imaging process is quite different from before.As is mentioned,multiple spectral image features are presented in the same frame of image at the same time,clearly and completely revealing the visible and invisible information of the observed scene and providing accurate relative comparison of the feature points of different bands in the same field of view Position,thus fundamentally overcome the shortcomings of traditional multi-device working methods,such as low efficiency,complicated use,and poor accuracy.With the intention to refine the quality of the fused image,specifically for the fused image and the processing procedure before image fusion,the image fundamental frequency detail separation technology through the nonlinear filter,the detail optimization technology of the adaptive Gaussian filter,the inverse histogram projection technology of the redundant grayscale and the adaptive gain control technology impressingly highlight the refined quality of the original blurred infrared images.For the sake of improving the noise and edge detail blur problems caused by traditional pyramid layered fusion,a new technology which implement algrithoms from the level of pixel.Using the good detail extraction ability of bilateral filtering combined with pyramid layering operation can maximize the acquisition of all the details of the image and make the fused image has pixel-level alignment capabilities and clear and excellent edge details of the target object.For the color fusion of the image,under the YUV color model,the pyramid layering model which fuse the reference figure and the unique new figure is implemented and a regional energy fusion method that compares the detail layer and the original low-quality color layer is designed.It should be noted that the final procedure highly refines the quality of the image by converting the blurred image to RGB color space,which correspondingly generate a unique multi-color image.This method makes the target highly impressive and prominent,which to some sense redefine the application of fusing methods due to its high quality and rich color range.Through the pixel-level fusion effect test,image detail enhancement resolution test,and color reconstruction objective index analysis,this paper proves that the target of the fusion result is more prominent,the colors are thus “indeed” natural compared with before,and it also proves that the fusing method inverstigated here lies in the level of pixel,which strengths the featuring analysis capabilities.Such a highly convenient and easy-to-control system not only reduces the difficulty and cost of application site deployment,but also improves the flexibility of user operations.It can be widely used in security early warning,border inspections,power patrols,equipment troubleshooting and other fields.