Landsat Remote Sensing Image Timing Change Detection Based On Deep Neural Network

Landsat data has been used to study how the earth’s surface has changed over the decades due to its long history and relatively high spatial resolution.Timely and accurate detection of the characteristics of the earth’s surface can help people understand the relationship and interaction between human and natural phenomena,and can promote better decision-making and development.Although there have been many successful cases in monitoring and detecting environmental changes,there are still great challenges in applying multi-time images to obtain timely information of the earth’s environment and human activities.It is worth noting,however,that significant progress has been made in overcoming technical barriers in recent years through the development of new platforms and sensors,and that the wide availability of large historical image archives makes long-term change detection and modeling possible.This development has prompted further research and development of more advanced image processing methods and new methods for processing time-series image data.Therefore,this paper will carry out relevant research with the help of deep neural network algorithm.Therefore,the research objectives of this paper are as follows:1.In this paper,a deep reinforcement learning optimization algorithm based on MDP as the core of the recurrent is proposed to automatically optimize the be optimized neural network,accelerate the neural network training process,and help the neural network jump out of the local optimum point.2.Considering that clouds and cloud shadows in Landsat remote sensing images affect the availability of Landsat remote sensing images,this paper proposes a cloud detection and removal method based on spatial semantic perception to realize end-toend automatic cloud detection and removal.3.In this paper,a time series change detection algorithm for Landsat remote sensing images based on multi-objective semantic segmentation neural network was proposed,which realized the simultaneous detection of multi-objective in Landsat remote sensing images.Compared with previous algorithms,this algorithm has excellent detection accuracy.