| Vegetation coverage is an intuitive index of soil erosion and is the data base of vegetation restoration in soil and water conservation.Vegetation restoration rate as an important indicator of evaluating the measure of soil erosion control,they are too time-consuming to finish evaluation task of large-scale projects within the administrative time.High-altitude remote sensing interpretation technology is the main method of regional vegetation research in agroforestry.However,it is not suitable for small-scale vegetation coverage study due to its large-area coverage of satellite remote sensing image.UAV low-altitude remote sensing technology has been used in land and resources statistics,agricultural monitoring,disaster investigation and many other fields.Once the technology is introduced into soil and water conservation,not only can it brings the high flexibility,real-time and image clarity of small-scale vegetation information survey into full play,but also can make up for the inefficiency of traditional methods.To improve the field verification efficiency and the accuracy of the evaluation of soil and water conservation work,this paper explores the application of UAV low altitude remote sensing technology in the calculation of vegetation coverage in small areas.Firstly,the spectral characteristics of vegetation and other typical features in UAV images are analyzed.Based on the regularity of vegetation area pixels:green band>red band>blue band,the vegetation index of RGRI,EXG,VDVI,NGBDI and NRGRDI are used to calculates all the pixels of the image so as to get the gray scale of each vegetation index distribution.According to the characteristics of the histogram of each vegetation index,the thresholds are respectively determined using the bimodal histogram method and the maximum threshold method.The parts which calculated values are larger than the threshold value are classified into vegetation,otherwise they are classified into non-vegetation.After that,classification results' accuracy are verification.Then,according to the classification results,the number of vegetation parts is divided by the total number of regional pixels to get the regional vegetation coverage.At last,the concept of vegetation coverage multi-temporal mapping is introduced,and the vegetation coverage of different monitoring periods is calculated by using two vegetation indices with the highest precision in order to explore the relationship between the correct rate of vegetation index and vegetation density.Research results1?The results showed that the total extraction accuracy of vegetation information was above 82%,and the vegetation coverage rate was similar to that of the reference value.The total correct rate of VDVI was 97.47%and the Kappa coefficient was 0.87,which means the results are highly consistent.The validation test's outcomes show that the vegetation index method is reliable and applicable.The experimental results also show that visible-band difference vegetation index(VDVI)is most suitable for the vegetation coverage estimation of UAV low altitude remote sensing images.2.After the analysis and comparison,this method has the highest accuracy when used in the high vegetation density period.Therefore,the optimal calculation method should be selected according to the vegetation growth status,so as to ensure the reliability of the monitoring results and make full use of the human resources.Main conclusionIn this paper,the method of calculating vegetation coverage based on UAV remote sensing image can greatly reduce the staff's outdoor workload and improve the objectiveness of the vegetation coverage results.It is a powerful complement to the traditional methods,which is in line with the automation,efficiency and informationize trend.The relationship between the vegetation index's accuracy and vegetation density studied in this paper gives reference for the rational use of UAV remote sensing technology in soil and water conservation,emphasizes the importance of the combination of traditional and new methods as well. |
PDF Full Text Request