Project: High Resolution Waferlevel reflowable EDoF Camera Module

Progresses in digital technologies created a revolution in the doCO of photography. _x000D_Nearly all cameras are digital, and the number of cameras is still exponentially growing. Also, many doCOs now use digital cameras: photography of course, but also medical applications, video monitoring, or manufacturing control. Progresses have been following the Moore law: the size of sensors has been dramatically decreasing and the computational capacity has been increasing in the same time. _x000D__x000D_Digital processing also allows performances that were simply impossible in classical photography, as for instance, having a a large enough depth of field with only 3 lenses at a fixed focus. _x000D__x000D_Recent advances in camera module manufacturing are waferlevel and reflowable camera modules that are much smaller and cheaper than traditional camera modules, because the camera is fixed on the sensor while this latter is manufactured by the SMT process (hence exposed high infra red radiation)._x000D_As for now, the technology applies only to small resolution (typically VGA or sub megapixel)._x000D__x000D_The resolution is still limited by several factors:_x000D_- the quality of optical design is not sufficient (limited number of available materials, difficulty to make several layers)_x000D_- the constitution of the camera module itself prevents the use of more complex optical systems as zooms or autofocus lens. _x000D_- furthermore, even for a fixed focus length, it is not possible to tune the focusing position once the module is manufactured. Because of inevitable manufacturing tolerances, this constraint is not compatible with high resolution sensors._x000D__x000D_DxO Labs has already introduced a revolutionary concept, called DxO Digital Optics, consisting in a co-design of the optical system of the camera along with the digital processing leading from the Raw sensor information to the generated image. _x000D_Heptagon is the worldwide leader in the manufacturing of waferlevel and reflowable lenses._x000D_The image processing research team in the University of Balear Islands is one of the best academic teams in the image signal processing algorithms. They have developed (among others) a denoising method which outperforms all the state of the art methods. _x000D__x000D_We believe that the upcoming “reflowable” and “wafer-level” lenses technologies combined with our know how acquired on our previous technologies (i.e. “DxO Digital Optics” technology) will allow to make a step forward in the miniaturization of the camera-modules. _x000D_We estimate that most of the COstream of big OEM will use this kind of technologies, for 2, 3 and 5 Mpixels cameras._x000D_In term of volume, this corresponds to approx 250M cameras/year, from mid 2009 to 2012._x000D__x000D_The project WAFLE aims to push the limits on three crucial aspects: _x000D_1. outstanding image quality, _x000D_2. much smaller systems, _x000D_3. reducing production cost while increasing productivity._x000D_ _x000D_The project includes two phases: _x000D__x000D_Phase 1: Conception of new optical design using reflowable waferlevel lenses. _x000D_Camera modules using reflowable lenses can be assembled on the sensor and the chip in a single manufacturing process called SMT. This implies that the lens can stand exposure to high infrared radiation with temperature in a range of 270°C. _x000D_The purpose is to manufacture smaller camera modules on the same production line as the sensor, in order to reduce both the size and the cost of the modules. Obviously, the new designs have to support high resolution sensors (at least 3Mpixel), contrarily to current designs whose resolution is typically VGA, because of the optical aberrations appearing at higher resolutions. _x000D_However, the WAFLE project will enable us to correct these aberrations and produce even better image quality. DxO Labs and Heptagon will design and manufacture waferlevel reflowable lenses. These lenses will then be mounted on a camera module by a subcontractor. _x000D__x000D_Phase 2: New algorithmic methods of image generation. _x000D_The CO concept of previous generation is that the digital processing is not only able to correct some flaws of the optical part of the camera, but also to nicely use some information provided by optical aberrations. In order to relax even more the optical constraints and get better quality images, new algorithms of image generation are needed. In particular, a new demosaicing method will be developed (by UIB jointly with DxO Labs), with a better rendering of thin details, noise removal and the correction of optical aberrations as chromatic aberrations. _x000D_The project will include the porting of these new algorithms on embedded platforms by DxO Labs.

Acronym	WAFLE (Reference Number: 4303)
Duration	01/05/2008 - 01/03/2010
Project Topic	The WAFLE project consists in the conception of new optical design using reflowable and waferlevel lenses which are compatible with high resolution camera (3 and 5 Mpix) and co-optimized digital processing, allowing low-light and/or extension of depth of field properties
Project Results (after finalisation)	The CO outcome of the project is the first ever camera module using a Wafer level lens with a 3MPix sensor with extended depth of field capability (EDoF). _x000D_Together with hardware developments, progress in image processing algorithms helped to increase the image quality. In particular, better demosaicing and better denoising had spectacular results on RAW images.
Network	Eurostars
Call	Eurostars Cut-Off 1

Project partner

Number	Name	Role	Country
3	DxO Labs	Coordinator	France
3	Heptagon	Partner	Switzerland
3	Universitat de les Illes Balears	Partner	Spain