Research On High Dynamic Range Imaging Technology Of Visible Light Camera

The natural light intensity ranges from the brightest star of 120000 Lux to the night sky starlight of 0.00035 Lux,with a brightness contrast of 109:1.The human eye can handle a contrast ratio of 105:1,while the contrast performance of a conventional camera is only 103:1.Therefore,the dynamic range of the image obtained by using a normal camera is often insufficient,and does not reflect the actual situation of the actual scene.The obtained image is either overexposed at an extremely bright place,causing the portion of the image to be saturated and losing information,or underexposure in an extremely dark place,resulting in this part of the image noise is mixed with the scene details and is blurred.Therefore,more advanced imaging equipment is required to achieve high dynamic range images.Currently,there are two ways to obtain high dynamic range images,one is hardware and the other is software.On the hardware side,more advanced sensors can be developed,such as increasing the full-trap charge capacity,reducing readout noise,etc.,and adding auxiliary circuits such as logarithmic compression,gamma compression circuits,etc.,to the back end of the image sensor.In software,most of them use multi-frame image fusion algorithm.The common multi-exposure fusion algorithm is to control the camera to capture images at different exposure times to obtain images of the same scene region under different brightness.Among them,the long exposure time image provides scene details in a low brightness environment,and the short exposure image provides scene details in a high brightness environment.Finally,using the weight function,the regions with the richest information in the sequence images of different exposures are extracted and merged,so that a high dynamic range image with good actual scene information is obtained.This paper is based on the research of image quality of space-based cameras.For space-based cameras,the spatial target material is complex,and the imaging environment is in space with less light source,so the brightness of the same scene target varies greatly,requiring a wider dynamic range.Combined with the above analysis,this paper mainly completed the following work.Firstly,the workflow of the image sensor is analyzed in principle,and the key parameters affecting imaging are introduced.The calculation methods are given for some of the parameters.Secondly,according to the environmental characteristics of the space target,the optical scattering characteristics are simulated.Combined with the bidirectional reflectance distribution function in radiometry,the target irradiance is analyzed.In the model of bidirectional reflection distribution function,the appropriate Davies model is selected.Then the typical material parameters of the space target are selected and simulated with a certain camera.The pixel values and noise distribution of the image under different exposure times are obtained.Then,according to the requirements of the project indicators,the theoretical design methods of two typical high dynamic range image sensors are analyzed.A method of using dual gain dual channel sampling to increase the dynamic range of the image,a method using multiple slopes to increase the dynamic range.Combined with a given optical system,the dynamic range of the two sensor theories is calculated and the specific exposure time at which the range is reached is calculated.Finally,a new high dynamic range image fusion algorithm is proposed.The algorithm is based on the multi-exposure Laplacian pyramid improvement,and introduces phase consistency as the weighting coefficient of edge extraction,which makes the extraction of edge details of images with different exposure time robust and does not change with the change of illumination.The simulation test based on standard image shows that the proposed algorithm is superior to the original algorithm in subjective display effect and objective evaluation index.At the same time,the indoor experiment was carried out by using the standard dynamic range test equipment.The dynamic range of the fused image achieved the expected effect,and the effectiveness of the algorithm was verified.