- Title: Motion Detection and Temporal Filtering of Noisy Image Sequences
- Abstract: When capturing digital video there is always some amount of noise in the re- sulting signal. This is most evident in low-light conditions. To reduce this noise we have evaluated several algorithms that separates noise from motion. Some of these algorithms include temporal ltering while others only detect motion and generate a motion mask. The generated motion masks are then used to deter- mine what parts of the image that can be ltered temporally by using weighted averaging. The algorithms are prototyped using Matlab and some are also im- plemented in OpenCL for testing in real-time using actual cameras.The methods have been analysed regarding the quality of the ltered output and the accuracy of the generated motion masks. This has been done using a synthetic image se- quence, where a noise-free reference exists. Real-world captured sequences have been judged subjectively.
- Start Date: March 12, 2012
- Supervisor: Karl Åström
- Supervisor: Magnus Oskarsson
- Student: Helena Olén, Marcus Wennermark
Questions: webmaster
Last update: 2013-04-11
Centre for Mathematical Sciences, Box 118, SE-22100, Lund. Telefon: +46 46-222 00 00 (vx)