Recognition of coal dust image based on improved differential evolution particle swarm optimization

To separate the individual particle from the coal dust image particle fusion region accurately and confirm the internal mechanism of image characteristic parameters,an effective model based on improved differential evolution par- ticle swarm optimization is proposed. It includes three steps as follows:firstly,the change rules of coal dust parameters are analyzed and image characteristic model is set up. After that,the particle swarm overlap region is determined. Sec- ondly,the relationship between image parameters and dust characteristics are derived and the intersection model is set up in the edge feature points. Finally,the particle fitness function values are determined and the global optimal position is obtained. The results show that the proposed characteristics rules yield more effective decision. The most interfering points can be eliminated and the computational burden is reduced significantly. Furthermore,the diversity of particle swarms is maintained effectively by adding mutation operator,so particles premature convergence is avoided and over- lapping particles are separated effectively. When the particle size is