Research And Implementation Of Video And Image Dehazing Algorithm Based On DM6467T

Images acquired in hazy environment always become blurred; have worse visibility and less detailed information, which cause great impact on traffic monitoring, aerial photography, intelligent navigation and other fields. Therefore it is of great theory and practical significance to study how to improve the quality of the image and obtain clear images.This dissertation introduces the de velopment status of dehazing algorithms and Da Vinci technology, describes Da Vinci technology principle and related hardware processor device in detail, introduces Da Vinci software package and user guide of Codec Engine, and presents the Da Vinci software development process. Da Vinci launched by TI Company is specially used for video processing. It is the appearance of this technology that makes engineers are liberated from cumbersome development process of underlying so urce code, and mainly focus on the development of video processing algorithm itself. It greatly reduces the workload of engineering staff and greatly shortens the project development cycle.In algorithm development, this dissertation introduces several image dehazing algorithms, and presents the characteristics of each algorithm. We mainly research dark channel prior dehazing algorithm, introduce the algorithm principle and process steps, and give the algorithm processing results. Then, improve and optimize the dark channel prior dehazing algorithm, and transplant the algorithm to DSP hardware system. On the premise of guarantee the quality of processing results, the optimized algorithm improves the efficiency of the algorithm and reduces the complexity. This dissertation also presents an image dehazing algorithm based on YUV space. This algorithm defogs mainly in the brightness space, and recovers the original clear image by using physical model of atmospheric attenuation, recover the clear image.This dissertation mainly studies the hardware programming implementation of embedded algorithm; therefore it also gives the hardware migration process of dehazing algorithms, and makes an explanation and description for each step of the algorithm transplantation in detail. This dissertation introduces the process of memory allocation as well as the cooperative communication between ARM and DSP in detail. Several related video formats and Da Vinci algorithm standard are also presented in this paper. Experimental results show that this algorithm can achieve a good effect on video and image dehazing in Da Vinci embedded DSP system, and has a certain practical value for engineering application. However, due to the complexity of the algorithm and the limits of hardware resource, it is need to further improve the real-time property of the algorithm.At the end of this dissertation, a review and summary to the full text is made, point out the shortcomings of present research and put forward the issues that need to improving and deeper researching. This paper has well reference significance on the same research fields.