Study On Underwater Bubble Detection Based On Optical Properties

Hydrothermal and cold seeps have become one of the hotspots in oceanology which opens up anew direction in geological, geochemical and geo-biological research. It provides potential solutionsfor global energy crisis as well as greenhouse gas environmental contaminations. Detection ofhydrothermal and cold seeps has been conventionally undertaken by inefficient close range chemicalsensing method. This research proposes a long range detection of cold seeps through the sensing ofgas bubbles emission from the cold seeps as a means to localize them. This research develops asensing method for gas bubbles and to visualize them through image processing using Matlab. Thedevelopment of this bubble detection technology is potentially important to other applications suchas detection of sea wakes and pipeline gas leaks under ocean etc.In this research the scattering property such as phase function, specula and diffuse scattering ofbubble within different media has been studied by simulation and the result is verified by experiments.Experimental set up includes a laser source as illuminant, an imaging camera and a photodetector assensor to detect the back scattering of light from the bubble. The angle of the reflection has beenadjusted during the experiment to optimize a better signal to noise ratio. Two mirrors have beendeployed to change the light path in order to reduce reflections from the glass of the water tank.Images of bubbles have been processed by two different methods to compare the results:(i) inthe static image analysis the second order Butterworth low-pass filter and grey level histogram havebeen utilized for image de-noising, image segmentation and image binning. The bubbles are thenmarked by the8-connection method and geometric properties of the bubbles, such as area, pixelposition coordinates and diameter of bubbles are evaluated. The result is saved into the data array.(ii) Dynamic image analysis: the velocity of bubbles has been measured in two ways. Animproved algorithm based on the combination of bubbles influence value and displacements havebeen proposed for a more effective way to track the bubble. This algorithm is easy, effective and hasobtained good matching. The other way is to detect the back scattering using a mirror with a centeredhole and together with a photodetector. The average velocity of bubbles can be calculated by fastFourier transformation.The results showed bubble detecting can be achieved by the experiment system, and thegeometric information and velocity of bubbles can be acquired accuracy during the experiment. Thisdetect system has some reference meaning on future application.