Research on seabed classification of marine ranching based on high-precision acoustic images

As the foundation for ecological engineering of marine ranching, artificial reefs play an important role in the restoration of marine ecological environment and the proliferation of fishery resources. With a large number of artificial reefs, the seabed of marine ranching has gradually transformed into highly heterogeneous benthic habitats, which directly affects the development and reproduction of bottom organisms. How to accurately identify seabed habitats has become an important issue in the benthic ecological research of marine ranching, and remote sensing of the seabed using acoustic technique combined or individually are providing solutions to this problem. Currently, the acoustic seabed classification technique has been applied to ecological research in many sea areas successfully, but it is rarely used in marine ranching. Also, the science of acoustic seabed classification is at its nascence, and there is still no widely accepted agreement about the standardization of instruments and methods. Focusing on the application of acoustic technique in seabed classification of marine ranching, we developed a new method of seabed telemetry and identification based on side scan sonar (SSS) and multi-beam echo sounder (MBES), and applied it in Tianjin Marine Ranching. The SSS and MBES acoustic images objectively reflect the hardness and topography of seabed respectively, which closely relates to seabed type. But most seabed types are not directly measurable using remote survey techniques. It is always necessary to measure the seabed types using in situ ground truthing methods, which can be linked in some way to the characteristics of acoustic images, and then the seabed types can be identified and classified by using acoustic images. In our research, 70 seabed in-situ samples were analyzed and connected with the corresponding acoustic images, and the results show that key image features to identify the seabed type were brightness and texture of SSS image and topography derived from MBES image. Based on the believable relationship between seabed type and acoustic image characteristic, the seabed type of Tianjin Marine Ranching was identified, classified and quantified. The results show that ① six seabed types were found in Tianjin Marine Ranching, including artificial reefs, oysters, mud, clam-mud, oyster-mud and shell-mud, and the accuracy of seabed classification method based on acoustic characteristics is 90%, which can meet the needs of benthic ecological research in marine ranching; ② the mud seabed has subsided to 58.13%, and the shellfish seabed has increased to 7.96%, which demonstrates that the shellfish proliferated in Tianjin Marine Ranching; ③ 0.266 km² of new oyster reef are closely surrounding artificial reefs, and 0.313 km² of oyster-mud seabed will be developed into oyster reef, which verifies the restoration effect of Tianjin Marine Ranching on living oyster reef. Compared with conventional benthic survey methods, our method of acoustic seabed classification based on high-precision acoustic images can avoid sampling errors, reduce human disturbance, discover accurate position of shellfish and improve the accuracy of seabed classification. This study explored the acoustic classification method suitable for the seabed of marine ranching, discussed the complement and perfection of acoustic technique to traditional benthic survey methods, clarified the close relationship between marine ranching seabed types and acoustic image characteristics, and verified the restoration effect marine ranching on oyster reefs. Our research results can be effectively applied to coastal habitat restoration, accurate assessment of shellfish resources, and the evolution mechanism of oyster reefs.