基于SAR和光学数据融合的山区森林干扰检测

doi:10.13466/j.cnki.lczyyj.2024.03.008

摘要/Abstract

摘要：

高效的森林干扰检测方式可以及时预防并减轻森林灾害,保护生态系统功能。针对融合多源数据对山区森林干扰进行检测时往往受到地形影响的问题,构建坡向分类算法,划分检测区域,以降低地形起伏对雷达变化比(RCR)的影响。通过融合合成孔径雷达(SAR)和光学卫星数据,利用改进的RCR方法与NDVI时间序列,提出一种新的森林干扰检测方法。研究结果表明:1)通过坡向分类算法改进RCR方法有效扩大了19.48%检测区域面积,可覆盖更多的干扰区域,并提升检测准确性;2)采用SAR与光学数据相融合的森林干扰检测方法,其总体检测精度为89.24%,与仅采用单一传感器的SAR数据、光学数据相比,检测精度分别提高了11.11%、13.32%。相较于单一传感器方法,此方法能在不同时间和天气条件下获取丰富的连续检测信息,在提高森林干扰的检测能力方面具有更大的潜力和优势,可为今后森林资源管理和生态保护提供更加全面和准确的信息支持。

关键词: 森林干扰, SAR, 雷达变化比, NDVI时间序列, 坡向分类

Abstract:

Efficient forest disturbance detection methods can prevent and mitigate forest disasters in time and protect the ecosystem.To address the issue of forest interference in mountainous areas,which is often affected by terrain when integrating multi-source data,this study develops a slope direction classification algorithm to delineate the detection area.This mitigates the effect of terrain relief on the radar rate of change(RCR).A novel forest disturbance detection method was proposed on the basis of the fusion of synthetic aperture radar(SAR)and optical satellite data,utilizing an enhanced RCR approach with NDVI time series.The results were as follows:1)The enhanced RCR methodology markedly expands the detection area 19.48% through the slope classification method,encompassing a greater scope of interference areas and enhancing the detection accuracy.2)The overall detection accuracy based on the fusion data of SAR and optical satellite is 89.24%,which is 11.11% and 13.32% higher than that of SAR and optical satellite with only a single sensor.Compared with the single-sensor method,this research method can obtain rich,continuous detection information under different time and weather conditions,and it has greater potential and advantages in improving the detection capability of forest disturbance,which can provide more comprehensive and accurate information support for forest resource management and ecological protection in the future.

Key words: forest disturbance, SAR, radar change ratio, NDVI time series, slope direction classification

中图分类号:

TP79
S771.8

王博, 陈永刚, 闫彦廷. 基于SAR和光学数据融合的山区森林干扰检测[J]. 林草资源研究, 2024,(3): 60-69.

WANG Bo, CHEN Yonggang, YAN Yanting. Forest Disturbance Detection in Mountainous Areas Based on SAR and Optical Data Fusion[J]. Forest and Grassland Resources Research, 2024,(3): 60-69.

图/表 10

图1

表1

表2

图2

图3

图4

表3

图5

表4

图6

参考文献 25

[1]	朱教君, 刘足根. 森林干扰生态研究[J]. 应用生态学报, 2004, 15(10):1703-1710.
[2]	梁建萍, 王爱民, 梁胜发. 干扰与森林更新[J]. 林业科学研究, 2002, 15(4):490-498.
[3]	张园, 陶萍, 梁世祥, 等. 无人机遥感在森林资源调查中的应用[J]. 西南林业大学学报, 2011, 31(3):49-53.
[4]	孙福洋, 李晓松, 李增元, 等. 近实时中高空间分辨率森林火灾监测系统展望[J]. 遥感学报, 2020, 24(5):543-549.
[5]	张丽云, 赵天忠, 夏朝宗, 等. 遥感变化检测技术在林业中的应用[J]. 世界林业研究, 2016, 29(2):44-48.
[6]	BROSOFSKE K D, FROESE R E, FALLKOWSKI M J, et al. A review of methods for mapping and prediction of inventory attributes for operational forest management[J]. Forest Science, 2013, 60(4):733-756.
[7]	SANNIER C, MCROBERTS R E, FICHET L V, et al. Using the regression estimator with Landsat data to estimate proportion forest cover and net proportion deforestation in Gabon[J]. Remote Sensing of Environment, 2014, 151(1):138-148.
[8]	杨辰, 沈润平, 郁达威, 等. 利用遥感指数时间序列轨迹监测森林扰动[J]. 遥感学报, 2013, 17(5):1246-1263.
[9]	KENNEDY R E, YANG Z, COHEN W B. Detecting trends in forest disturbance and recovery using yearly Landsat time series:1.LandTrendr—Temporal segmentation algorithms[J]. Remote Sensing of Environment, 2010, 114(12):2897-2910.
[10]	ZHU Z, WOODCOCK C E, OLOFSSON P. Continuous monitoring of forest disturbance using all available Landsat imagery[J]. Remote Sensing of Environment, 2012, 122(1):75-91.
[11]	HANSEN M C, KRYLOV A, TYUKAVINA A, et al. Humid tropical forest disturbance alerts using Landsat data[J]. Environmental Research Letters, 2016, 11(3):034008.
[12]	REICHE J, VERBESSELT J, HOEKMAN D, et al. Fusing Landsat and SAR time series to detect deforestation in the tropics[J]. Remote Sensing of Environment, 2015, 156(1):276-293.
[13]	饶月明, 王川, 黄华国. 联合多源遥感数据监测四川木里县森林火灾[J]. 遥感学报, 2020, 24(5):559-570.
[14]	冯琦, 陈尔学, 李文梅, 等. 基于ALOS PALSAR数据的热带森林制图技术研究[J]. 遥感技术与应用, 2012, 27(3):436-442.
[15]	MERMOZ S, TOAN L. Forest disturbances and regrowth assessment using ALOS PALSAR Data from 2007 to 2010 in Vietnam,Cambodia and Lao PDR[J]. Remote Sensing,2016, 8(3):217.
[16]	BOUVET A, MERMOZ S, BALLÈRE M, et al. Use of the SAR shadowing effect for deforestation detection with Sentinel-1 Time Series[J]. Remote Sensing, 2018, 10(8):1250.
[17]	REICHE J, HAMUNYELA E, VERBESSELT J, et al. Improving near-real time deforestation monitoring in tropical dry forests by combining dense Sentinel-1 time series with Landsat and ALOS-2 PALSAR-2[J]. Remote Sensing of Environment, 2018, 204(1):147-161.
[18]	COLSON D, PETROPOULOS G P, FERENTINOS K P. Exploring the potential of Sentinels-1 & 2 of the copernicus mission in support of rapid and cost-effective wildfire assessment[J]. International Journal of Applied Earth Observation and Geoinformation, 2018, 73(1):262-276.
[19]	REICHE J, MULLISSA A, SLAGTER B, et al. Forest disturbance alerts for the Congo Basin using Sentinel-1[J]. Environmental Research Letters, 2021, 16(2):024005.
[20]	RÜETSCHI M, SMALL D, WASER L T. Rapid detection of windthrows using Sentinel-1 C-Band SAR data[J]. Remote Sensing, 2019, 11(2):115.
[21]	SHIMIZU K, OTA T, MIZOUE N. Detecting forest changes using dense Landsat 8 and Sentinel-1 time series data in tropical seasonal forests[J]. Remote Sensing, 2019, 11(16):1899.
[22]	杭州市临安区人民政府. 临安概况[EB/OL]. 2022-12-12[2024-03-10]. https://www.linan.gov.cn/col/col1366287/index.html.
[23]	罗时雨, 童玲, 陈彦. 全极化SAR图像的山地低矮植被区域土壤含水量估计[J]. 遥感学报, 2017, 21(6):907-916.
[24]	ALMEIDA D R A d, BROADBENT E N, FERREIRA M P, et al. Monitoring restored tropical forest diversity and structure through UAV-borne hyperspectral and lidar fusion[J]. Remote Sensing of Environment, 2021, 264(1):112582.
[25]	KOSKINEN J T, PULLIAINEN J T, HALLIKAINEN M T. The use of ERS-1 SAR data in snow melt monitoring[J]. IEEE Transactions on Geoscience and Remote Sensing, 1997, 35(3):601-610.

成像时间	拍摄卫星	产品类型	极化方式
2021-11-14	Sentinel-1A	SLC	VV&VH
2021-11-26	Sentinel-1A	SLC	VV&VH
2021-12-08	Sentinel-1A	SLC	VV&VH
2021-12-20	Sentinel-1A	SLC	VV&VH
2022-01-01	Sentinel-1A	SLC	VV&VH
2022-01-25	Sentinel-1A	SLC	VV&VH
2022-02-06	Sentinel-1A	SLC	VV&VH
2022-03-14	Sentinel-1A	SLC	VV&VH
2022-03-26	Sentinel-1A	SLC	VV&VH

成像时间	拍摄卫星	条带号	云量/%
2021-11-14	Sentinel-2A	N9999_R089	0.01
2021-11-24	Sentinel-2A	N0500_R089	1.04
2021-12-04	Sentinel-2A	N0500_R089	0.70
2021-12-19	Sentinel-2B	N0301_R089	0.69
2021-12-29	Sentinel-2B	N0301_R089	0.38
2022-01-03	Sentinel-2A	N0301_R089	0.00
2022-02-27	Sentinel-2B	N0400_R089	1.24
2022-03-14	Sentinel-2A	N0400_R089	4.74

分类方法	OA/%	F1/%	UA/%	PA/%	可检测面积/km²	占比/%
坡度分类	76.04	71.92	84.38	62.67	1 124.64	41.20
坡向分类	78.13	76.67	83.47	70.89	1 656.42	60.68

区域	SAR融合权重	光学融合权重	AUC值
向坡	0.50	0.50	0.813 4
	0.55	0.45	0.799 6
	0.60	0.40	0.792 1
	0.65	0.35	0.827 6
	0.70	0.30	0.861 7
	0.75	0.25	0.851 3
	0.80	0.20	0.765 8
	0.85	0.15	0.752 2
	0.90	0.10	0.741 4
	0.95	0.05	0.729 7
背坡	0.50	0.50	0.813 4
	0.45	0.55	0.806 5
	0.40	0.60	0.823 1
	0.35	0.65	0.830 6
	0.30	0.70	0.823 5
	0.25	0.75	0.801 2
	0.20	0.80	0.762 1
	0.15	0.85	0.750 5
	0.10	0.90	0.736 5
	0.05	0.95	0.723 4