Integrating multimedia and artifical intelligence for pest prediction and aeration control of stored grain bins

In the present study, an intelligent system with human-machine interface of knowledge acquisition is established to implement the pest prediction and the aeration control of stored grain bins. In the system, recurrent neuro-fuzzy network models are proposed to predict temperature evolvement in the grain bins. 3D multimedia displays of node sensor-measured temperatures, and its gradient distributions of the given grain layer and their variations in the interval of the given time are used to extract the system knowledge in the current and future. The results of the experiment in the two grain depots in northeastern China have verified the effectiveness of the system.