Research On Data Compression Method Of Forest Ecological Network Based On Compressed Sensing

Forest resources are one of the most important and irreplaceable renewable resources on the planet.Protecting the safety of forest ecological environment is an important prerequisite for the development of human civilization.Establishing a micro-environment monitoring station in the forest area is an effective measure to protect the forest ecological environment.At present,micro-environment monitoring stations deployed in forest areas have energy shortages.Most of the forest micro-environment monitoring stations upload the original data directly to the cloud server,resulting in a large energy consumption of the monitoring station.In the long-term operation process,there is a problem of insufficient data supply resulting in data loss.Therefore,how to reduce the energy consumption of data acquisition at the monitoring station but ensure the accuracy of data acquisition at the monitoring station is the primary problem in current research.In this paper,a data acquisition method based on compressed sensing for forest ecological network is proposed.The main research contents are as follows:(1)The method of forest micro-environment data collection based on fixed dictionary is studied.The original data is compressed and analyzed by using the most common fixed dictionary such as DCT,DFT,etc.,and the sparse dictionary is evaluated by comparing the sparseness error,the reconstruction error,and the data compression ratio.The results show that the data compression method using DCT dictionary is slightly better than DFT dictionary in error performance,but the compression ratio of DFT dictionary is much higher than DCT dictionary.(2)The method of forest micro-environment data collection based on learning dictionary is studied.The K-SVD dictionary learns the original eigenvalues in the original data,so it can be adaptively sparsely represented by the original data.The results show that the K-SVD dictionary data compression method is better than the DFT dictionary as a whole,but there are also a small number of samples that are more suitable for DFT dictionary.(3)A forest area micro-environment data collection method based on switching dictionary is proposed.According to the different representation of the original data structure,the method of function simulation is used to classify the data,and the concept of switching factor is proposed.Different types of data are respectively compressed and reconstructed using the most suitable dictionary.The results show that the data compression method using the switch dictionary in the forest micro-environment monitoring station has smaller reconstruction error and higher compression ratio than the single dictionary.