| Bamboo is one of the most important forestry resources with great ecological and economic value.The distribution information of bamboo forest can be extracted from Forestry land efficiently by using remote sensing methodsatlow cost.In this study,the maximum likelihood method and support vector machine method were used to extract the distribution information of bamboo forest in Jian'ou City,Fujian Province in 2017 based on Landsat8OLI image,resulting in a thematic map of bamboo forest in the study area.hen soil samplings were taken from the bamboo forest and analyzed so as to get a better understanding of the characteristics of Organic Matter,Total Potassium and pH in bamboo forest soil.The steps to extract bamboo from forestry includes image preprocessing,optimal band screening and band fusion,and supervised classification based on different classification methods.Firstly,based on the ground control point in the pre-processing operation,the image is geometrically refined,which effectively improves the accuracy of the image.Then using the OIF index method combined with the spectral curves of the ground objects drawn by the field survey,the ray image common band gray value and the bamboo forest extraction band combination were analyzed,and the three band groups 356,256 and 367 were determined to be suitable for bamboo forest extraction.Starting from image color and classification identifiability,it is determined that the 356 band is the best band combination for extracting bamboo forest.Based on the actual situation of the study area,the maximum likelihood classification method and the support vector machine(SVM)method are used to supervise and classify the bamboo forest,other tree species,water bodies and other four types of features in the 356 band.The results of the confusion matrix accuracy verification show that the maximum likelihood classification Kappa coefficient(85.30%)is lower than the SVM classification method(88.78%),and SVM method satisfies the classification requirements and the result is the best classification result.Finally,a map of the bamboo forest in Jianye City in 2017 was drawn.Referring to the bamboo forest thematic map,a five-point sampling method is adopted on the bamboo forest soil in the study area,and sampling points are Identified in the bamboo forest thematic map and the Soil samples were collected.The soil samples of the bamboo growth layer(0-30 cm)were analyzed in the laboratory,and the three properties most closely related to bamboo growth were obtained,namely organic matter,total potassium and pH.In order to study the statistical and spatial characteristics of the three attributes,conventional statistics,geostatistical analysis and Kriging interpolation analysis were performed on the three attributes.The results showed that the average value of organic matter was 3.95%,the average value of total potassium was 19.06 mg/g,and the average value of pH was 4.94<7.00.The soil of bamboo forest was generally acidic.The difference between the three values and the spatial difference are significant,among which the spatial difference of organic matter is the most significant.The organic matter content is mainly concentrated between 0.8710-1.2190(%),accounting for 72.32%of the total area.The difference in total potassium space is less than that of organic matter,and the highest potassium content is only 6%different from the lowest area.The pH is mainly concentrated between 4.5 and 5.0,which accounts for 55.37%of the total bamboo forest area.Bamboo forest soil with a pH of less than 4.5 is mainly concentrated in the southwestern part of the study area.In addition,the study also found that the spatial variability dominant factors of organic matter,total potassium and pH are different.pH is mainly affected by its own structural factors.Total potassium is mainly affected by random factors,while organic matter is subject to its own structural and random factors. |
PDF Full Text Request