Mike,

I'm afraid you will not find an acceptable solution. Definitely not something that could run on Raspberry Pi.

To suppress motion blur, you need to locally estimate PSF of the motion blur and do deconvolution. It is not a simple problem and the quality of the image reconstruction will get worse with noise and moving objects in the scene. On a mobile robot, you could use accelerometers to guide the motion estimation. If you are using a camera with rolling shutter, you will have to deal with that as well (e.g. http://graphics.stanford.edu/papers/stabilization/).

An interesting and simple to implement deconvolution method is presented in: James Gregson et al.; Stochastic Deconvolution, CVPR 2013. http://www.cs.ubc.ca/labs/imager/tr/2013/StochasticDeconvolution/

Have a look at http://www.cse.cuhk.edu.hk/~leojia/projects/motion_deblurring/index.html.

While the method may be beyond the processing power of the Pi, I recommend you take a look at the following paper.  It is one solution (not necessarily perfect) to your problem and has some discussion on other methods. 

Removing Motion Blur With Space–Time Processing Hiroyuki Takeda, Member, IEEE, and Peyman Milanfar, Fellow, IEEE

http://users.soe.ucsc.edu/~milanfar/publications/journal/MotionDeblurringFinal.pdf

Thanks guys for the info. I'll take a look at the suggested pages/papers. I have noticed that the picamera offers a shutter_speed() python call (http://picamera.readthedocs.org/en/latest/api.html#picamera.PiCamera.shutter_speed) that allows the adjustment of the camera's shutter speed (based on the view's illumination). If I can set the shutter speed faster would that help with blurring?

Likely it will help however, shutter speed is a function of available light and sensitivity of the sensor.  For a given camera, if you want to take a picture in low light, the shutter will need to stay open longer than if you take the picture in bright sunlight.  In any case, if the subject (or sensor) moves while the shutter is open you will get blurring.  If you reduce the amount of time the shutter is open by reducing shutter speed below its "ideal" setting for the amount of light, you will get less blurring but you will also get a darker image.  

"shutter speed is a function of available light and sensitivity of the sensor" , if enough light (illumination as it is referred to in the picamera docs) is available, the shutter speed can be sped up (made faster)? The picamera is capable of pretty impressive resolution (http://picamera.readthedocs.org/en/latest/fov.html#camera-modes). I'm collecting HD (1920x1080) images at 5 fps so I figured that I should be able to collect images without the blur.

Seeing how the final product I'm trying to capture is road imagery in poor conditions (weather, unmapped, unmarked, etc) I don't think it matters if the images come back darker, as long as they capture a realistic scene. I'm collecting now (summer) to establish a baseline as the weather is typically great with clear sky’s, and any available markers (both natural and man-made) show up well. I can then later do the same data collect runs during the winter and see how they compare.

Approximating the motion blur and then compensating for it, can be a reasonable approach.