Browse Prior Art Database

High Dynamic Range Capture Techniques using Multiple Exposures and Optical Flare

IP.com Disclosure Number: IPCOM000214243D
Publication Date: 2012-Jan-20

Publishing Venue

The IP.com Prior Art Database

Related Documents

Ref. 1 of D11110: OTHER

Abstract

High dynamic range imaging typically involves time sequential capture of multiple photographs. This cannot be easily applied to images with moving objects, especially if the motions are complex. In this paper we provide a two step process to get around this. Firstly, we optically encode both the low dynamic range portion of the scene and highlight information into a low dynamic image that can be captured with a conventional image sensor. This is achieved using a cross screen or star filter. Then we decode with software, both the low dynamic range image that can be captured with a conventional image and the highlight information. These two steps can then be combined to form an image of a higher dynamic range, than the regular sensor dynamic range.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 12% of the total text.

High Dynamic Range Capture Techniques using Multiple Exposures and Optical Flare

The idea in this disclosure allows the capture of a much more dynamic range within only two exposures. This could apply to professional and consumer HDR video capture.

Abstract

High dynamic range imaging typically involves time sequential capture of multiple photographs. This cannot be easily applied to images with moving objects, especially if the motions are complex.

In this paper we provide a two step process to get around this. Firstly, we optically encode both the low dynamic range portion of the scene and highlight information into a low dynamic image that can be captured with a conventional image sensor. This is achieved using a cross screen or star filter.

Then we decode with software, both the low dynamic range image that can be captured with a conventional image and the highlight information. These two steps can then be combined to form an image of a higher dynamic range, than the regular sensor dynamic range.

Background

Camera sensors can capture a certain maximum number of photons before they start to saturate and no longer register additional light. Although it is impossible to increase the saturation point by increasing the capacity of the sensor electron, producing large sensors is excessively expensive and reduces sensor resolution.  Such sensors are also hard to justify for general imaging applications because on average, only a small portion of a scene contains very bright spots and thus needs high capacity sensors.

The human visual system has developed a clever mechanism to cope with highly saturated scene regions, such as highlights or light sources. Like camera sensors, the photoreceptors in the human retina are also prone to saturate. However, the visual system is able to infer higher brightness of those saturated regions from glare, which is produced by the light that is scattered in the ocular fluid and spread over the retina. The glare surrounding bright areas boosts their perceived brightness, giving additional information to the brain that this part of the scene is much more than the photoreceptor saturation point. (See fig 1, Chap 1, pg 289. ‘Glare Encoding of High Dynamic Image Ranges.’)

Previous Trials

Previously we disclosed the idea of bracketing two exposures from a 3D camera to capture HDR, using an ND filter or differing exposures. However, the approach only allows for the capture of two exposures which must have some overlap. This limits the dynamic range that can be captured. Many real-world scenarios contain much higher dynamic range then can be captured with this technique. This invention attempts to redress this problem.

We have produced a method to spatially disperse bright features or highlights in a scene that would have otherwise saturated a sensor. The trick is to use an optical filter which causes flare with a characteristic pattern. Spatially spreading the light limits the amount that the sensor is saturated, th...