Abstract:Traffic noise map play a significant role in urban environmental noise management. In order to reduce the prediction error for big size noise mapping project, a measurement based inverse approach for traffic noise sources was presented, and the detailed method of noise map correction calculation was introduced step by step. The approach improved correction solving efficiency by mixing original noise mapping calculation result and inverse sources effect, avoided modify the prediction model directly, and ensured that the calculation results in target area can match the noise propagation rules defined by prediction model. Then, the approach mentioned above was implemented as a software package on the noise mapping platform developed by our group. At last, a real traffic noise mapping project in Beijing was introduced to demonstrate the method. By comparing the 6 groups of experimental result, it was showed that the correction error can reach less than 1.1dB on the location of measurement points. On the other points in calculation area, the accuracy of sound levels was improved more or less. The correction error was related to two factors measurement points: the contribution rate of main noise source and the influence area of measurement points. The correction error can reached less than 2.5dB on the points located in influence area of measurement point. The experimental results revealed that this method was suitable to inverse the traffic sources by measurements, generate correction calculation result and improve solving quality of noise map.