Color Restoration for Underwater Object Detection and Tracking

button3

Color Restoration for Underwater Object Detection and Tracking

By: Mohd Faid Bin Yahya (PhD Student)

 

Underwater robots have been one of the most actively undergoing researches in the world as it is the pinnacle subject in the field of robotics and ocean engineering. Some of the applications which utilizes underwater robot are underwater monitoring, seabed mapping, and underwater structure inspections. The ability for the underwater robots to autonomously interact with the underwater environment is crucial for the success of aforementioned applications. One of the most important aspects which prove to be a very challenging issue is the usage of underwater vision. The reason being is due to the limited detection ranges and the blurry underwater visions. Being able to use vision when a robot is in underwater will open up new underwater application opportunities such as in sensing and close range detections.

In order to overcome the drawbacks of using vision in underwater, the underwater color restoration algorithm had been introduced in [1]. The algorithm has been conducted on an underwater robot. Also, the algorithm has been tested in the real time image processing scenarios. Figure 1 shows the underwater target objects where the algorithms had been tested to differentiate and tracked all of them at once.

 

3a mac15

Figure 1. Underwater target objects.

 

Accordingly, Figure 2 below shows that the algorithm successfully differentiate as well as track between the cross,
circle, cone, and cylindrical shape objects.

3b mac15

Figure 2. Target object tracking with labels 1,2,3, and 4 in the images denote cross, cone, sphere, and cylinder, respectively.

 

For future works, other forms of objects can be tested and further improvement can be made by implementing it into a real autonomous underwater vehicle.

 

Reference(s):

  1. L. Donghwa, K. Gonyup. K. Donghoon, M. Hyun, and C. Hyun-Taek. (2012). Vision-based object detection and tracking for autonomous navigation of underwater robots. Ocean Engineering48(2012)59-68.