Low level image processing is the base of localization. The paper discusses time domain filtering and frequency domain filtering according to the noise disturbance in the surrounding images.

Filtering algorithms base on space, frequency, wavelet and time have been introduced to reduce noises caused by environmental and lighting conditions in the video image sequences.

If you feed the output from your test rig into a real-time computer filtering software and mount everything to a solid base, you would have a system as capable as any available.