Pipeline recognition in side scan sonar image using adaptive network based fuzzy inference system (ANFIS) classifier

In applications of Autonomous Underwater Vehicle Navigation the automatic detection of underwater objects in sonar images is the most important part. A process of analyzing multi-beam sonar images for localization and underwater object detection is a new framework presented in this research. The main idea of the project is based on Pipeline Detection. Morphological filtration method is used for the more efficient form of Denoising and segmentation of the image. Image quality assessment features and image quality enhancement features such as GMSD, PSNR, SSIM, MSE, along with Bicubic interpolation and super resolution is also implemented for better image classification. The image is been preprocessed with filtration technique but also with image segmentation process such as Fuzzy-C means segmentation and further been sent to Adaptive Neuro Fuzzy Interference System (ANFIS). Classification which has good generalisation ability to accurately spot pipeline. The proposed framework which can figure out the position accurately and detect objects is been experimented by exploiting the image processing technique.