Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator
In recent years, machine vision algorithms have played an influential role as core technologies in several practical applications, such as surveillance, autonomous driving, and object recognition/localization. However, as almost all such algorithms are applicable to clear weather conditions, their p...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-10-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/20/5795 |
| _version_ | 1797551060671266816 |
|---|---|
| author | Dat Ngo Seungmin Lee Gi-Dong Lee Bongsoon Kang |
| author_facet | Dat Ngo Seungmin Lee Gi-Dong Lee Bongsoon Kang |
| author_sort | Dat Ngo |
| collection | DOAJ |
| description | In recent years, machine vision algorithms have played an influential role as core technologies in several practical applications, such as surveillance, autonomous driving, and object recognition/localization. However, as almost all such algorithms are applicable to clear weather conditions, their performance is severely affected by any atmospheric turbidity. Several image visibility restoration algorithms have been proposed to address this issue, and they have proven to be a highly efficient solution. This paper proposes a novel method to recover clear images from degraded ones. To this end, the proposed algorithm uses a supervised machine learning-based technique to estimate the pixel-wise extinction coefficients of the transmission medium and a novel compensation scheme to rectify the post-dehazing false enlargement of white objects. Also, a corresponding hardware accelerator implemented on a Field Programmable Gate Array chip is in order for facilitating real-time processing, a critical requirement of practical camera-based systems. Experimental results on both synthetic and real image datasets verified the proposed method’s superiority over existing benchmark approaches. Furthermore, the hardware synthesis results revealed that the accelerator exhibits a processing rate of nearly 271.67 Mpixel/s, enabling it to process 4K videos at 30.7 frames per second in real time. |
| first_indexed | 2024-03-10T15:39:23Z |
| format | Article |
| id | doaj.art-5375a47065ff4152a5c0e7a953b0c4e4 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T15:39:23Z |
| publishDate | 2020-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-5375a47065ff4152a5c0e7a953b0c4e42023-11-20T16:56:52ZengMDPI AGSensors1424-82202020-10-012020579510.3390/s20205795Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware AcceleratorDat Ngo0Seungmin Lee1Gi-Dong Lee2Bongsoon Kang3Department of Electronics Engineering, Dong-A University, Busan 49315, KoreaDepartment of Electronics Engineering, Dong-A University, Busan 49315, KoreaDepartment of Electronics Engineering, Dong-A University, Busan 49315, KoreaDepartment of Electronics Engineering, Dong-A University, Busan 49315, KoreaIn recent years, machine vision algorithms have played an influential role as core technologies in several practical applications, such as surveillance, autonomous driving, and object recognition/localization. However, as almost all such algorithms are applicable to clear weather conditions, their performance is severely affected by any atmospheric turbidity. Several image visibility restoration algorithms have been proposed to address this issue, and they have proven to be a highly efficient solution. This paper proposes a novel method to recover clear images from degraded ones. To this end, the proposed algorithm uses a supervised machine learning-based technique to estimate the pixel-wise extinction coefficients of the transmission medium and a novel compensation scheme to rectify the post-dehazing false enlargement of white objects. Also, a corresponding hardware accelerator implemented on a Field Programmable Gate Array chip is in order for facilitating real-time processing, a critical requirement of practical camera-based systems. Experimental results on both synthetic and real image datasets verified the proposed method’s superiority over existing benchmark approaches. Furthermore, the hardware synthesis results revealed that the accelerator exhibits a processing rate of nearly 271.67 Mpixel/s, enabling it to process 4K videos at 30.7 frames per second in real time.https://www.mdpi.com/1424-8220/20/20/5795haze removalmachine learningsupervised learninghardware acceleratorfield programmable gate array |
| spellingShingle | Dat Ngo Seungmin Lee Gi-Dong Lee Bongsoon Kang Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator Sensors haze removal machine learning supervised learning hardware accelerator field programmable gate array |
| title | Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator |
| title_full | Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator |
| title_fullStr | Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator |
| title_full_unstemmed | Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator |
| title_short | Single-Image Visibility Restoration: A Machine Learning Approach and Its 4K-Capable Hardware Accelerator |
| title_sort | single image visibility restoration a machine learning approach and its 4k capable hardware accelerator |
| topic | haze removal machine learning supervised learning hardware accelerator field programmable gate array |
| url | https://www.mdpi.com/1424-8220/20/20/5795 |
| work_keys_str_mv | AT datngo singleimagevisibilityrestorationamachinelearningapproachandits4kcapablehardwareaccelerator AT seungminlee singleimagevisibilityrestorationamachinelearningapproachandits4kcapablehardwareaccelerator AT gidonglee singleimagevisibilityrestorationamachinelearningapproachandits4kcapablehardwareaccelerator AT bongsoonkang singleimagevisibilityrestorationamachinelearningapproachandits4kcapablehardwareaccelerator |