Bei Zhang

In the scenario of planning auxiliary decision-making, traditional map symbol labeling methods have problems such as low efficiency and insufficient accuracy, which are difficult to meet the needs of real-time planning. To address this pain point, this paper proposes a dual-branch network model based on MobileNetV2 to realize the synchronous and accurate recognition of the category and angle of hand-drawn point symbols, and then automatically match the recognized standard point symbols to the corresponding positions on the map. The research collects 174 types of hand-drawn point symbol samples and adopts a step-by-step training strategy: first pre-training the category recognition branch, then freezing the parameters of the MobileNetV2 feature extraction network, and only updating the parameters of the fully connected layer to improve the generalization ability of the model. Experimental results show that the model achieves a Top1 accuracy of 97.8% and a Top3 accuracy of 99.7% in the point symbol category recognition task, and a Top1 accuracy of 95.8% and a Top3 accuracy of 99.7% in the angle recognition task, which meets the practical application needs of planning. This research provides efficient and accurate human-computer interaction support for the new type of planning system, and promotes the landing application of multi-modal human-computer interaction technology.

The ship information system has the characteristics of complex operation and large amount of information. Due to the limited cognitive ability of the operator, the overload of cognitive resources will reduce the efficiency of the task completion of the complex information system. Existing studies have provided a set of indicators to characterize the change of operational status of maritime personnel, but there is a lack of specific research on the sensitivity and operability of each indicator. In this context, based on the indicators representing the change of operational status of maritime personnel, this study adopts the form of questionnaire to carry out the sensitivity experiment of indicators. In the experiment, the first and second indexes were scored respectively, and the experimental results showed that performance and subjective scores were obviously better than physiological data measurement methods. In terms of physiological data, the scores of eye movement, electrocardiogram and facial expression were better than those of electrodermal activity and electroencephalogram. The conclusion of this study can be used to select the operational status change index of ship information system operators.