| Forest plays an irreplaceable role in regulating the global carbon balance and the global climate.One of the effective feasible solution to forestry application bottlenecks is using wireless sensor network to monitore Forest ecological indicators. However, the data analysis and processing may add certain difficulties because of the intensive deployment of the sensor nodes, the enormous quantity of sensor data, and data loss and redundancy; A large number of node layout may be lead to a waste of resources.In this paper, considering the non-negative of the sensor data,takes the method of matrix non-negative low-rank approximation(NNLRA), combined with MATLAB programming to reduce the humidity data dimension that gathered from wireless sensor networks and provides a new thought to the processing of compressing the data of carbon emissions and monitoring of carbon sinks. At the same time, the factor matrix resulted from the method optimizes the selection of the sensor network nodes allocation. It is of great significance to the optimum allocation of sensor networks and sensor network information collection, etc.The main work of this paper are as follows:(1) The NNLRA improves the calculation of singular value eigenvectors in singular value decomposition in processing sensor data and leads to a relatively good dimension reduction. It not only cuts down the memory space but also speed up the calculation.(2) Compared to PCA, the relative error between dimension reduction data and original data is smaller with the method of NNLRA, the value computed by NNLRA is1.009%.(3) Using nonnegative low-rank decomposition, we get non-negative matrix. Through the non-negtive matrix and the original sensor data matrix, we get the best node layout scheme of the sample polt.In this paper, we combine the LRA with NMF, given a non-negative low rank approximation (NNLRA) algorithm, realized the sensor data matrix dimension reduction, saves storage space, and also accelerate the speed of the algorithm; the nodes arrangement number reduced from80to49.It realizes the optimization of wireless sensor network of monitoring of carbon sinks and save resources. |
