" Speckle noise reduction in TV holography fringes using wavelet thresholding" by Guillermo H. Kaufmann ; Gustavo E. Galizzi seems to be interesting

http://opticalengineering.spiedigitallibrary.org/article.aspx?articleid=1074033

I would recommend you to go for Golay coded excitation. It has good speckle noise elimination property. 

FYI:

http://www.medicalimageanalysisjournal.com/article/S1361-8415(12)00055-2/abstract

Concerning the acquisition stage, if your start point is a long coherence length laser (gas, laser diode), you can use a RGG (Rotating Ground Glass) diffusor.

Article Partial coherence effects in digital holographic microscopy ...

I agree with Luis. You can do some averaging using a diffusor

It depends on your design. If one design a hologram with periodic array instead of a single phase distribution, no speckle would happen.

Zheng, Can you explain what you mean and give some relevant references?

I would recommend the paper

F. Yaras, H. Kang, and L. Onural, "Real-time phase-only color holographic video display system using LED illumination," Appl. Opt.  48, H48-H53 (2009).

I used LED to illuminate sinusoidal phase grating recorded as an analog thin hologram or generated by a SLM to produce sinusoidal fringes with reduced speckle noise for optical metrology.

Regards,

Elena Stoykova 

Vladimir Farber, I guess you can understand it with the attached file (part of holographic image).

A speckle noise reduction is a hard problem. I can recommend a time average. Many years ago I received an impressive result using a laser illumination through a cuvette with milk. I suppose that a liquid crystal can be used instead of milk.  

To reduce speckle noise in the reconstruction of a hologram, we use a spinning diffuser and a sheet of holographic diffuser. The holographic sheet diffusers are specified by angles. We use either the 10 degree or 15 degree holographic diffuser. The spinning diffuser was originally a spinning aperture, bought commercially. We cut a disc of sheet protector film to make the 5 degree spinning diffuser. The combination of a static and a spinning diffuser completely eliminates the speckle noise.

You might also see the full reference text  "Chapter: Vibration Measurement by Speckle Interferometry between High Spatial and High Temporal Resolution" on my researchgate page.