Through questionnaire investigation, statistics is obtained from major Chinese ocean shipping companies regarding their relevant staff's opinions on power station operation training. Principles of fuzzy comprehensive evaluation method is introduced and the structural system for assessment indices is also established with analytic hierarchy process to conduct fuzzy evaluation on teaching staff's quality, training courses, training textbooks, the combination of theory with practice, training duration and method of teaching etc. The weighting of the evaluation indices is determined by its average value decided by fuzzy entropy, evaluation model is established to get solved, empirical study is carried out on the present training strategy, training result is reasonably evaluated and training measures are improved.

Human factor is main cause for vessel safety. In order to reduce the risk of the human factor risk, it is necessary to consider the human factor scientifically and study the human effect on the navigation security. According to seafarers' natural condition, body condition, work pressure and tired condition, the model based on Markov Chain is established in this paper. The probability of mental state is calculated on base of statistics, and the assessment model is presented. In this paper, internal-external locus of control scale were examined in COSCO seafarers, the results by simulation indicates that the seafarers reaction time is reduced rapidly, that the hidden Markov tree method presented by authors can effectively improve the flexibility, reliability and the safety of the ship sailing at sea.

Human behavior is a major factor in the occurrence of accidents at shipping companies. Thus, safety training is essential to improving workers' safety awareness and behavior, constructing a corporation safety culture, and improving economic efficiency. For keeping diversity and combination between multinational shipping enterprise cultures and their influence on organizational behavior and individual behavior, the article analyzes a method of establishing safety training system and builds a training system model based on the PDCA. In addition, this study establishes the basic guidelines of operation and protection mechanisms, as well as other various elements, to identify the best way to conduct safety training in the future.

In the ship navigation safety chain, the seaman's factor is very important for a ship's maneuvering safety. A structure of seaman-ship integrated model (SSIM) was established, and related data being used was proposed. The important function of the method was briefly discussed. The application of SSIM to the predication of a ship's maneuvering was conducted to obtain an excellent ships maneuvering quality, and the seaman-ship optimal control model was found by simulation study. The simulation results show that the obtained optimal seaman-ship control characteristic can fully take the advantages of ships maneuvering technical performance. The proposed SSIM plays an extremely vital role in ship navigation security.

How to reduce maritime traffic accidents resulting from human errors which is an important factor of causing marine accidents is one of the main concerns. According to the characteristics of ships at sea and the crew airworthiness theory, this paper establishes ship management model on the base of circuit theory. Then it simulates the ship organization, the environment, the influence that the crew's own resistance to the airworthiness of the crew and the crew airworthiness output under company's incentive. It also provides theoretical basis for crew seaworthiness management and effective scheduling.

Ship detection over long distances is crucial for the visual perception of intelligent ships. AI techniques, particularly machine learning and deep learning, have achieved a series of breakthroughs in this field. However, due to the limited pixels of ships over long distances, such objects often suffer from weak feature representation and are susceptible to interference in complex environments. To address these challenges, this paper proposes an improved architecture named Ship-DiffDet, based on DiffusionDet. First, we redesign the backbone feature extraction network and propose IDC-Net, which utilizes inception depthwise convolution to enhance feature extraction efficiency while reducing computational complexity. Second, to tackle the difficulty of effectively extracting features from small objects, we design a Hybrid Pooling Attention-enhanced Feature Pyramid Network. By incorporating a hybrid pooling attention mechanism, it strengthens multi-scale feature fusion, thereby improving the performance of the detection heads. Furthermore, we introduce a multi-order gated aggregation mechanism into the dynamic detection head to optimize dynamic convolution and further compress the model’s parameter count. Experimental results demonstrate our method achieves an effective balance between detection accuracy and computational efficiency. On our custom-built small-object ship dataset, the proposed method improves AP₅₀ by 1.7% over the baseline while reducing the parameter and FLOPs counts by 48.8% and 22%, respectively.