| Autor: |
Pranav Wani, Kashif Usmani, Gokul Krishnan, Bahram Javidi |
| Jazyk: |
angličtina |
| Rok vydání: |
2024 |
| Předmět: |
|
| Zdroj: |
IEEE Access, Vol 12, Pp 166643-166651 (2024) |
| Druh dokumentu: |
article |
| ISSN: |
2169-3536 |
| DOI: |
10.1109/ACCESS.2024.3493601 |
| Popis: |
This paper presents an experimental analysis of information loss in degraded environments for traditional 2D imaging and 3D integral imaging. We consider mutual information to quantify this information loss. Our experimental analysis shows that 3D integral imaging preserves more scene information in degraded environments considered here than traditional 2D imaging. Additionally, as an example of the effects of information loss, we experimentally analyze the performance of integral imaging-based object depth localization in foggy environments. We compare the performance of three commonly used integral imaging based depth localization methods, that is, mutual information, minimum variance, and maximum voting strategy in foggy environments. For this purpose, we use illustrative laboratory scenes recorded in varying fog levels with and without partial occlusions. We assume the availability of bounding boxes corresponding to each object. An increase in fog severity results in increased information loss as measured by mutual information. Our analysis shows that all three algorithms perform comparably in clear environments and can localize an object’s depth with good accuracy. The depth localization accuracy decreases in light to moderate foggy environments. However, mutual information provides more accurate depth information for light to moderate foggy environments compared to the other two methods. All three algorithms fail to provide reliable depth information for severe foggy environments. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
