Real-time Video Defogging,Compression And Transmission Based On Embedded System

For nighttime foggy weather,the atmospheric light intensity in the scene is relatively weak and contains an artificial light source.At this time,the main light source is an artificial light source,and the atmospheric light value of the foggy image is no longer a fixed value,but the atmospheric light function changed by different artificial light source positions,for this reason,this paper proposes an improved nighttime real-time defogging algorithm based on Gaussian filtering and information loss theory constraints.The image processing quality of this algorithm is significantly improved,and the clarity and contrast of the image are effectively improved.The upper optimization algorithm achieves a quick defogging effect with high real-time performance.At the same time,for high-definition camera shooting video in a foggy scene,the picture quality is reduced,the amount of image data is large,and the problem of real-time defogging processing and network transmission is required.This paper optimizes the night video image proposed in this paper with the core real-time defogging algorithm,uses the H.264 standard to compress and encode video data,and transmits video data through a UDP multicast protocol network.A multi-task binding multi-core processor parallel processing algorithm is used to establish a real-time night video collection,defogging,compression.and the network transmission system.The system performance is superior,both high quality and high stability characteristics.The main research work of this paper is as follows:1.This paper studies and analyzes the defogging principle and experimental results of various classical image dehazing algorithms by learning the fog imaging characteristics of nighttime foggy scenes,and establishes an atmospheric transfer model suitable for nighttime scenes with artificial light sources,based on Gaussian filter function estimation.The atmospheric light function,which changes with the position of the light source,adjusts the atmospheric transmission transmittance in the foggy image based on the information loss theory,and adopts a fast median filter to refine the atmospheric transmission transmittance,and proposes an improved nighttime real-time defogging algorithm to achieve rapid and efficient Defogging the fog image at night.2.This article adopts the i.MX6 Quad quad-core processor chip with Cortex-A9 architecture.In the software development platform built with the Linux kernel as the operating system,it captures video image data by driving and using the OV5640 high definition image sensor device and optimizes the dehazing algorithm based on Gaussian filtering and information loss theory,uses H.264 format for compression encoding operation of video data by the hardware coding unit,and transmits the compressed video data through the UDP multicast protocol for network transmission,and finally completes at the host computer.real-time display.3.This article uses a multi-core processor parallel algorithm in the embedded system to bind multiple tasks such as video data acquisition,fast image defogging,VPU video compression encoding,and UDP network transmission to different cores of the processor.Utilize system resources to improve the overall performance of embedded systems.Based on the defogging algorithm and embedded system proposed in this paper,the simulation experiment is performed.Compared with the experimental results of various classic and improved dehazing algorithms,this algorithm has a higher degree of recovery from foggy images,the image color is natural after processing,the defogging effect is better and this algorithm can be applied to the embedded system efficiently and in real time.