A Biomimetic Model of Adaptive Contrast Vision Enhancement from Mantis Shrimp

Mantis shrimp have complex visual sensors, and thus, they have both color vision and polarization vision, and are adept at using polarization information for visual tasks, such as finding prey. In addition, mantis shrimp, almost unique among animals, can perform three-axis eye movements, such as pit...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Sensors (Basel, Switzerland) Switzerland), 2020-08, Vol.20 (16), p.4588
Hauptverfasser: Zhong, Binbin, Wang, Xin, Gan, Xin, Yang, Tian, Gao, Jun
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Mantis shrimp have complex visual sensors, and thus, they have both color vision and polarization vision, and are adept at using polarization information for visual tasks, such as finding prey. In addition, mantis shrimp, almost unique among animals, can perform three-axis eye movements, such as pitch, yaw, and roll. With this behavior, polarization contrast in their field of view can be adjusted in real time. Inspired by this, we propose a bionic model that can adaptively enhance contrast vision. In this model, a pixel array is used to simulate a compound eye array, and the angle of polarization (AoP) is used as an adjustment mechanism. The polarization information is pre-processed by adjusting the direction of the photosensitive axis point-to-point. Experiments were performed around scenes where the color of the target and the background were similar, or the visibility of the target was low. The influence of the pre-processing model on traditional feature components of polarized light was analyzed. The results show that the model can effectively improve the contrast between the object and the background in the AoP image, enhance the significance of the object, and have important research significance for applications, such as contrast-based object detection.
ISSN:1424-8220
1424-8220
DOI:10.3390/s20164588