APPLICATION FOR 3D SCENE UNDERSTANDING IN DETECTING DISCHARGE OF DOMESTICWASTE ALONG COMPLEX URBAN RIVERS

摘要: In our study we use 3D scene understanding to detect the discharge of domestic solid waste along an urban river. Solid found Ciliwung River in neighbourhoods Bukit Duri and Kampung Melayu may be attributed households. This is part due inadequate municipal infrastructure services which has caused those living river rely upon it for disposal. However, there been little research understand prevalence household Our aim develop a methodology that deploys low cost sensor identify point source using image classification methods. To demonstrate this describe following five-step method: 1) strip GoPro images are captured photogrammetrically processed dense cloud generation; 2) depth each generated through backward projection clouds; 3) supervised method based on Random Forest classifier applied view dependent red, green, blue (RGB-D) data; 4) locations can then mapped by projecting classified 5) landscape elements into five types, such as vegetation, human settlement, soil, water waste. While work still ongoing, initial results have demonstrated possible perform quantitative studies help reveal estimate amount present bank.