Research On Enhancement Algorithm For Night Image

With the rapid development of electronic technology,some professional or portable camcorders,cameras and monitoring equipments has been widely used in our daily,such as traffic monitoring,medical support,video forensics,night driving,film production and so on.Image quality is a very important index to measure the performance of these devices.However,due to the insufficient or uneven illumination at night,images captured by related equipments typically have low visibility and contrast,which can result in deterioration of the perceived quality and the loss of the information.Furthermore,these images are not convenient for observation and analysis,which will have a negative impact on related applications,especially in image matching,fusion,analysis,detection and segmentation.Therefore,night image enhancement has become a hot and challenging research topic in the field of image processing.The main purpose of night image enhancement is to improve the image's brightness and contrast,and to make details visualization.In the processing,many methods directly process the original image,however,in this paper,we propose a novel approach for night image enhancement based on the local background image.The method enhances the original image by adjusting the local background image.To avoid color cast and color distortion due,we firstly convert the image to the HSV color space,then process the V component.As much as possible to extend the enhancement effect,we use the local background image generated by the V as a processing object.In order to improve the image brightness and contrast,we have selected the appropriate transform function to extend low pixels range of the local background image and compress high pixels range of the local background image.For the problem that some part is too bright or too dark after the transformation,we optimize and improve transform function to minimize this phenomenon.Finally,we do the deconvolution processing in enhanced local background image to obtain enhanced V image.Experimental results show that the proposed method is very effective.Compared to a few traditional methods,this method ave very good results on some evaluation standards.