Noise Prediction Using LIDAR 3D Point Data - Determination of Terrain Parameters for Modelling

Abstract

AbstractNoise prediction at a location necessitates topographical data (positional information about the building, road, etc.), noise data (of sources), and a prediction model to forecast noise levels. This is important for setting up urban planning for noise pollution management. Variability in outdoor conditions of grounds along with the locations of noise sources and non-sources (i.e., noise receivers) make it difficult to predict noise levels accurately. Good noise prediction models need to determine related factors accurately, i.e., how noise can transmit from noise source to receiver positions- directly or indirectly. Existing techniques for noise modelling suffer due to approximation in use of technique for generation of terrain parameters along with the use of inadequate qualities of terrain data. Use of highly accurate and dense LIDAR data is planned to overcome deficiencies in data quality. Detail source to receiver noise propagation paths is tried to be determined. While propagating from noise source, it can directly reach to the receiver or reach indirectly after diffracting around barriers and buildings. Further, it can reflect from the ground and wall before reaching to receiver location. In the paper a technique is described using LIDAR data to generate terrain parameters. LIDAR 3D point cloud data are used as terrain data. An algorithm is proposed to determine all the possible paths in 3D from the noise source to the receiver. A point-to-point rigorous routing system capable of operating in 3D is created, specifically designed to suit the sonic propagation concept (unlike well know shortest path determination algorithms used primarily in 2D). Once the detailed path information is obtained, the detailed terrain parameters (distance, difference, reflection contribution, ground absorption etc.) are determined. These are then applied over semi-empirical noise model to determine the noise levels at receivers (or unknown locations). From the LIDAR data of RGIPT campus, building coordinates are extracted and then by applying the new routing algorithm it was possible to identify the paths, and extract terrain parameters. These are then applied over noise model to predict the noise map for RGIPT campus. Accurate prediction is ensured through use of accurate terrain data and efficient extraction of terrain parameters from detailed paths in 3D (unlike the conventional 2D shortest routing algorithms). Accurate management schemes. prediction scheme can help in developing efficient noise. KeywordsNoise modelingNoise mappingNoise predictionSound propagationTraffic noiseLIDAR