Information Compression Of Forest Microenvironment Based On Adaptive Switching Dictionary

The material rich nature is the ideal home for survival.As an inexhaustible natural resource,forest resources have the functions of flood storage,drought resistance,air purification,etc.in human production and life,there are a large number of plants and microorganisms in forest areas,which also provide important raw materials for industrial and agricultural production.However,although the total amount of forest resources continues to increase in China,there are still insufficient coverage and uneven distribution.It is extremely unfavorable for the sustainable development of the forest itself to have plant life,death and frequent fire disasters.Ecological monitoring of the forest microenvironment is a measure that can not only effectively judge the growth of plants,but also timely carry out fire risk early warning.Forest micro environment monitoring judges the plant growth environment through sensors,and uploads a large number of information data to the corresponding database of the detection platform.In the process of increasing the number of monitoring stations,data transmission needs to consume more energy and storage space.In remote forest areas,problems such as power failure and data interruption often occur due to the untimely energy supply.Data compression can improve these problems.In this paper,an adaptive switching dictionary method is proposed to study and analyze the sparse steps of the compressed sensing theory.Two sparse methods are compared: DCT dictionary and K-SVD dictionary in terms of compression performance and error.On the basis of reconstruction error,the samples are classified,and the adaptive switching dictionary is composed,so that different samples can automatically switch to the sparse ones with smaller error Law.In the selection of switching methods,two adaptive switching methods are used: BP neural network and SVM support vector machine.The adaptive switching dictionary is used to sparse and reconstruct the forest microenvironment information.The advantages and disadvantages of the two adaptive switching dictionaries are analyzed and compared,and the power consumption is measured in the actual GPRS transmission.The experimental results show that with the increase of sparsity,the sparsity errors of both fixed dictionaries tend to decrease gradually,but the decreasing trend becomes slower.For K-SVD dictionary,increasing the dictionary dimension will reduce the sparsity error,and the decreasing trend will gradually slow down.When a single sparse method is used,the K-SVD dictionary is used as the sparse method which can input training samples for training,and the effect is better than the traditional DCT sparse method.When the number of observation points is 60,the average reconstruction error is reduced by about 2%.The combination of the two dictionaries can effectively compress the data.By adding labels to the samples,the compression method has better adaptability to a single sample,and the average reconstruction error compared with DCT dictionary is reduced by about 3%.In the actual transmission process,the energy consumption of the transmission module for data transmission is more than twice that of the standby state.After data compression and transmission using the adaptive switching dictionary,the storage space is reduced by more than 50%,and the transmission power consumption is reduced by 80% when the number of observation points is 60.