Research On Laser Speckle Vibration Detection Technology Based On Vision

In the field of computer vision and image processing,research on video de-raining algorithms has more and more practical significance.In real life,driving on a highway under severe weather such as rain or heavy rain,due to the rain line blocking,makes driving difficult and dangerous.In addition,important captured images may blur or even weaken the behavior of criminals,and in traffic accidents,the criminal responsibility of the perpetrator cannot be identified.Furthermore,video rain removal is of great significance to the development of AI technologies such as target detection and recognition and manual driving.In this paper,we consider the distinguishable prior information of rainy video,clean video and rainline,and propose two video rain removal models based on tensor convex optimization.1.The video rain removal model based on the main direction of the rainline and the truncated nuclear norm.First,the first-order directional derivative filter is used to determine the direction of the rainline,and then the prior information of the rainline and clean video is analyzed;then,the truncated nuclear norm is used Instead of the tensor kernel norm,the low-rank of the video is guaranteed;then a model is built,and for this model,the model is solved by the alternating direction multiplier method;finally,the algorithm is verified by testing the video and comparing the algorithm.2.Based on the prior information of the direction gradient and the sparse group tensor vide to remove rain.First,analyze the directional gradients of the rainline and the clean video in the time dimension and the space dimension respectively in the time direction,the vertical direction and the horizontal direction,and then use the total variation regular term of a single direction to characterize the rainline and the clean video.The video has discriminative prior information;next,analyze the sparseness of the rain line,and use the group sparseness to describe the rain line to ensure the structured information of the rain line;then,build the video rain removal model,and then based on the characteristics of the model,design the Zhang The efficient solution algorithm of the quantitative optimization model,the analysis of the convergence of the algorithm,and the use of data experiments to verify.Finally,experiment with simulated data and real data,and then compare and verify our algorithm with several other excellent algorithms.