The diversity and complexity of sediment particle image is the main bottleneck restricting river sediment image segmentation algorithm to establish. By comparing the data of basic particle swarm optimization (PSO) algorithm, a modified particle swarm optimization (MPSO) algorithm and chaos particle swarm optimization (CPSO) algorithm in the application of sediment particle image, this thesis puts forward a kind of sediment particle image adaptive chaotic particle swarm optimization (ACPSO)algorithm. In this algorithm，the chaotic sequence is introduced to improve the local search ability of algorithm ,and at the same time, the algorithm dynamically adjusts weight factor and the variance of the population's fitness. The construction of Intelligent Transportation Systems (ITS) occupies a crucial position in the current wave of smart city. Effective and efficiency ITS needs two important conditions: plenty of traffic data and effective means of data analysis. Multi-source, heterogeneous, vague, uncertain traffic data fusion and sharing is the focus and difficulty of current research and application of ITS. The granular computing demonstrates a unique advantage in the information analysis and processing of massive, vague, uncertain and incomplete data. In this paper, we study the traffic information granular computing theory and build traffic information fusion model, framework and implementation program based on granular computing. We raise uncertainty reduction algorithms for traffic flow prediction and congestion recognition algorithms based on granular computing theory, which will provide new ideas and methods in the complex decision making under uncertainty problems of the transportation systems.