| Abstract Submission No. | ABS-1.1-0071 |
| Title of Abstract | Side Scan Sonar Image Classification for Mapping Sand Mining Features in the Chapora River, Goa |
| Authors | Gowthami C*, Sriram G, Lasya Priya, Komal Rani, Dewangan P |
| Organisation | CSIR-National Institute of Oceanography, Dona Paula, Goa 403004 |
| Address | viwambarapuram,main street,narsipuram,parvathipuram manyam,andharpradesh PARVATHIPURAM, andhrapradesh, India Pincode: 535522 E-mail: gowthamichappa1@gmail.com |
| Country | India |
| Presentation | Poster |
| Abstract | Sand mining in riverine systems poses significant geomorphic and ecological challenges, necessitating advanced methods for monitoring and impact assessment. As part of environmental and geomorphological investigations in the Chapora River, Goa, high-resolution side scan sonar (SSS), single-beam bathymetry, seismic profiling, and sediment sampling were undertaken along ~26-line km of the channel. The side scan sonar mosaics revealed distinct geomorphic and anthropogenic features, including sand ripple marks, sand bars, exposed rock outcrops, bridge pillars, and zones of active sand extraction. Notably, excavation imprints attributable to two distinct mining practices Type I (linear trenches), Type II (circular pits), debris, cohesive sediments and river banks were observed, reflecting variations in extraction techniques and their morphological footprints. To improve feature identification, pixel-based image classification was performed using ilastik, an interactive learning and segmentation toolkit. Originally developed for microscopic imaging, ilastik was adapted for acoustic imagery to automatically classify and map seafloor features. This approach enabled systematic categorization of multiple classes, including sand ripple fields, rocky outcrops, structural obstructions (bridge pillars), active mining zones (Types I and II), and additional geomorphic elements. The integration of automated classification with geophysical data enhanced the accuracy of riverbed mapping and facilitated objective delineation of mining-affected zones. The study demonstrates the potential of coupling geophysical surveys with advanced image segmentation techniques for riverine resource mapping. By providing spatially explicit datasets on sand distribution, mining footprints, and associated features, this work contributes to baseline inventories for environmental monitoring, sustainable sand resource management, and mitigation planning for anthropogenic impacts in fragile river ecosystems. |
| Keywords | Sand Mining, Side Scan Sonar, Image Classification |