Research On Fishing Vessel Trajectory Mining And Application Based On Big Data And Deep Learning

The use of large-scale fishing vessel trajectory data to control fishing vessels is helpful to strengthen the supervision of fishing vessels,and has important research significance for ensuring the sustainable development of marine fishery resources and the safety of fishing production.However,the traditional fishing vessel supervision method based on trajectory big data has problems such as untimely data processing and insufficient mining,so it cannot effectively supervise fishing vessels,which is not conducive to the healthy development of fishery resources and the protection of fishermen’s life and property safety.Therefore,this paper is mainly based on big data and deep learning technology to process the fishing boat trajectory data,conduct indepth research on the two key issues of fishing boat operation mode identification and fishing boat trajectory anomaly detection,design and implement a big data platform for dynamic monitoring of fishing boats,in order to achieve The purpose of intelligent fishing vessel supervision is to help strengthen the conservation of fishery resources and ensure the safety of fishery production.The main research content of this paper is as follows:(1)Research on identification of fishing vessel operation mode.In view of the large amount of fishing boat trajectory data and the problems of insufficient data information extraction and low recognition accuracy in the current identification methods of fishing boat operation methods,this paper uses Spark RDD to process fishing boat trajectory data in parallel,and proposes a method based on one dimensional convolution Convolutional Neural Network(CNN)and Gated Recurrent Unit(GRU)fishing boat operation pattern recognition model(1DCNN-SAGRU).The model uses one-dimensional CNN and GRU to fully extract the local spatial features and timing dependencies of fishing boat trajectory data,and introduces a self-attention mechanism to strengthen the model’s ability to focus on important information.Finally,the Dropout method and RAdam optimizer are introduced to optimize the model.optimization.The experimental analysis shows that the identification model of fishing boat operation mode can increase the recognition accuracy by up to 4.4 percentage points,indicating that the model can more accurately identify fishing boat trawling,purse seine and gillnet operations.(2)Research on real-time anomaly detection of fishing boat trajectories.The fishing boat’s own factors and changes in the surrounding marine environment may lead to abnormal fishing boat trajectories,threatening the safety of fishermen’s lives and property.Therefore,this paper fuses the characteristics of fishing boat trajectory data,fishing boat operation methods,and marine environmental data,and proposes a method based on Spark and DA-Seq2 Seq Model Fishing Vessel Trajectory Real-Time Anomaly Detection Algorithm(FVTRAD).The algorithm first improves the Seq2 Seq structure by Dual Attention(DA)and designs the fishing vessel trajectory prediction model(DA-Seq2Seq),then analyzes the predicted value and the real value and compares them according to the set threshold to detect abnormal trajectories,and finally introduces Spark The Streaming real-time computing framework improves the timeliness of fishing boat trajectory anomaly detection.The experimental analysis shows that the detection rate of fishing boat trajectory anomalies can reach 86.6%,and the false detection rate is only 2.5%.At the same time,it can also meet the real-time requirements when faced with large-scale and high concurrent trajectory data.It shows that the algorithm can more timely and effectively detect fishing boats with abnormal trajectories.(3)Design and implementation of a big data platform for dynamic monitoring of fishing vessels.According to the actual project requirements,in order to solve the problem of real-time receiving,processing and mining analysis of fishing boat trajectories under large-scale and high concurrency conditions,this paper designs and implements a big data platform for dynamic monitoring of fishing boats.The platform is mainly based on big data technologies such as Spark and HBase to receive and store fishing boat trajectory data and marine environment data in real time,and then apply the designed fishing boat operation mode recognition and fishing boat trajectory realtime anomaly detection algorithm to the platform,combined with Spring Boot,Vue And other related development technologies to realize functional modules such as fishing boat operation mode control and fishing boat track anomaly detection,and finally achieve the purpose of assisting users in decision-making and strengthening fishing boat supervision capabilities.