Mining And Application Of Multi-source And Multi-temporal Spatio-temporal Data In Coastal Zone

Coastal zone is the transition zone between land and sea,and is an important place where human activities and natural ecosystems interact.The multi-source and multi-temporal spatiotemporal data of the coastal zone play an important role in the scientific research and sustainable development of the coastal zone.It is of great scientific value and practical significance to effectively mine these spatio-temporal data through certain means,dig out the hidden information,and summarize the laws.Among them,the mining and analysis of coastline changes is a popular research direction.Most of the existing mining schemes are directly mining spatio-temporal data,and the analysis and mining of coastline changes requires us to extract the hidden information of coastline distribution,and then Perform change analysis.However,the existing coastline extraction schemes have poor adaptability and low precision.Therefore,this paper uses the spatio-temporal data such as multi-source and multi-temporal remote sensing images to design a spatio-temporal data mining scheme for coastal zones based on cluster segmentation and edge detection.First of all,this paper selects multi-source,multi-temporal high-resolution remote sensing images in the same area as the original data,and improves the traditional K-means algorithm by using batch training subsets,and realizes the clustering and segmentation of remote sensing images on land and sea.At the same time,the clustering time efficiency of the algorithm is improved.Then,using the clustering results,by improving and constructing the Canny edge detection algorithm,including designing a new median filter to replace the traditional Gaussian filter,increasing the direction of gradient calculation and introducing an adaptive algorithm,the extraction of the coastline in the image is realized.,while improving the extraction accuracy of the shoreline.Then,using the extracted results,the changes and causes of the coastline in the study area were analyzed by superimposing the base map,selecting observation points,and unifying the scale.Finally,according to the discussion in this paper,a prototype system for analyzing and mining coastline changes in coastal space-time data is designed and implemented,and the system test results meet expectations.The experimental verification of the research method used in this paper shows that compared with the average clustering time of the traditional K-means algorithm of 48.41 seconds,the average clustering time after the improvement is shortened to 30.49 seconds,and the time efficiency is increased by 37.02%.The traditional Canny edge detection algorithm was improved.After the extraction was completed,the coastline extraction accuracy in each year before the improvement was 76% on average by setting pixel checkpoints.After the improvement,it reached 83.56%,and the accuracy increased by nearly 7.5 percentage points.According to the analysis of the changes,the coastline of the study area has advanced about786 meters eastward on average in eight years,with an average annual advance of 98.25 meters.The experiments show that the coastal zone spatio-temporal data mining method used in this paper is effective for the extraction and change analysis of the coastline in the coastal zone.It provides an effective basis for development and coastal zone planning.