We propose a novel method of improving algorithms recognizing traffic lights in video sequences. Our focus is on algorithms for applications which notify the driver of a light in sight. Many existing methods process images in the recording separately. Our method bases on the observation that real-life videos depict underlying continuous processes. We named our method FSA (Frame Sequence Analyzed). It is applicable for any underlying algorithm and improves it by adding an additional result post-processing step. Our experiments are based on improving a published realtime traffic light recognition algorithm. Its general description has been provided by its authors, which allowed us to create a best-effort implementation for testing. We verify the effectiveness of the FSA method on a public dataset, acquiring very good results - improving the underlying algorithm in terms of all considered error measures. In the end, conclusions and possible future improvements are discussed.
Authors
- Dr inż Adam Blokus,
- prof. dr hab. inż. Henryk Krawczyk link open in new tab
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1109/iwssip.2018.8439187
- Category
- Aktywność konferencyjna
- Type
- materiały konferencyjne indeksowane w Web of Science
- Language
- angielski
- Publication year
- 2018
