Remote sensing technology is a comprehensive technology that detects the electromagnetic wave radiation information on the earth's surface through sensors from different working platforms far from the ground, and then obtains the geometric and physical properties of surface objects through information transmission, processing, interpretation and analysis. As a space exploration technology, remote sensing is a comprehensive sensing technology gradually formed with the development of aerospace technology and computer technology on the basis of aerial photography and interpretation, and has gone through three stages: ground remote sensing, aerial remote sensing and aerospace remote sensing.

Forestry remote sensing technology refers to the remote sensing technology that can be applied to forestry. It conducts the survey and record of forestry biology, geography, ecology and other relevant information through non-contact and non-field observation technology. China's forestry remote sensing technology research has gone through three important stages, from local application research, to basic high-tech research, and then to large-scale application.

At present, forestry remote sensing technology has been widely used in forest resource survey, forest resource monitoring, forest pest control monitoring, forest information extraction, forest fire monitoring, wildlife monitoring and forestry law enforcement. On the basis of consulting the representative literature and data on the research and application of China's forestry remote sensing technology, this paper combs and summarizes the development process and application status of China's forestry remote sensing technology.

1 The development of forestry remote sensing technology

The beginning of China's forestry remote sensing can be traced back to the 1950s, and has been developing for nearly 70 years so far. From 1951 to 1980, China's forestry remote sensing technology was in the stage of visual interpretation and application based on aerial photographs. During the past 30 years, the application of forestry remote sensing technology was limited to forest resource survey, as well as the application of partial forest fire and pest prevention and control.

In 1953, the aerial inspection team set up by the Ministry of Forestry of China used 1/250000 small-scale old topographic maps and medium-sized aircraft to carry out aerial survey in the basin; In 1954, China introduced the forest survey technology of the Soviet Union, and established a forest aerial survey brigade. The aerial photograph was used to prepare the mosaic sketch map and the sketch duplicate map. The aerial photograph was used for stereo interpretation, and the sub-classes of different land use types and forest types were divided, and a small low-speed aircraft was used to conduct a visual survey of the sub-classes of the interpretation division one by one in the air; In 1977, the forest resources inventory was carried out in Tibet by using the dual sampling method and satellite image mapping technology of the combination of the United States Land Satellite MMS image with the aerial photograph and the ground sampling measurement. So far, China has completed the forest resources inventory work by satellite remote sensing for the first time.

In the early 1980s, China established a remote sensing experimental field, imported software and hardware equipment for remote sensing data processing, started the research and development of a computer-based remote sensing data digitization processing and analysis system, and carried out a series of applied experimental studies such as forest spectrum, image processing of MMS data, spectral determination of tree species, land classification and information extraction, stock estimation, urban green space distribution and its impact on the environment, and forest fire monitoring, the research and application of forestry remote sensing technology has also been gradually diversified from a single forest resource survey.

With the introduction of the concept of "digital earth" in 1998 and the launch of the first high spatial resolution satellite IKONOS in 1999, the development of China's forestry remote sensing technology has also entered a fast lane, and the period from 1981 to 2000 is also known as the pioneering and innovative stage of satellite remote sensing. During this period, with the joint efforts of more than 140 scientific and technological personnel in China, the technical specifications for the comprehensive remote sensing survey of renewable resources were formulated for the first time, and major breakthroughs were made in the key technical fields such as information source evaluation, professional remote sensing survey, remote sensing series mapping, remote sensing image processing, and ecological benefit evaluation, so as to realize the analysis, management and prediction of multiple resource data.

After 2000, China has carried out a lot of research work on forestry remote sensing technology, including canopy information extraction, spaceborne hyperspectral remote sensing data preprocessing, forest type remote sensing identification method, sub-class division method based on SPOT5 image, forest leaf area index and canopy density estimation, airborne hyperspectral data dominant tree species identification technology, forest chlorophyll content inversion, crop species identification and growth detection. The research on forest pest control and other aspects has achieved fruitful results, forming a comprehensive monitoring technology system for forestry, establishing an integrated monitoring platform for forest resources, a comprehensive application service platform for forestry remote sensing, and the forestry remote sensing technology has been deeply and widely applied.

2 Application of forestry remote sensing technology

2.1 Forest resources survey

China's forest resources survey is divided into national forest resources continuous inventory, county-level unit forest resources survey and operational survey based on forest management operation design. Among the above three different types of forest resources survey, only the operation survey based on forest management operation design can be completed without the use of remote sensing technology. In the process of forest resources investigation, remote sensing (RS) technology is commonly used in combination with geographic information system (GIS) and global positioning system (GPS).

RS has the ability of real-time, fast and dynamic acquisition of spatial data, providing timely, accurate and large-scale forest resources data for GIS. GIS collects, stores, processes and analyzes data, and combines GPS to provide important means and auxiliary data for the extraction, analysis and application of remote sensing information, so as to improve the classification accuracy, storage capacity and data processing efficiency of remote sensing data.

GPS determines the longitude, latitude and elevation of any place on the earth's surface in real time and continuously, and provides accurate geographical location data for GIS. Combined with digital elevation model (Digital Elevation Model,DEM) and GIS, after coordinate transformation and image correction, the ground can be accurately registered with satellite or UAV remote sensing images to form a data model.

At present, TM satellite images are generally used to carry out national and large-scale forest resources survey, and SPOT or Quickbird images are used to carry out regional forest resources survey. In the remote sensing technology based on satellite images, because some measurement indicators require high resolution, the resolution of satellite images is difficult to meet the requirements.

In recent years, with the rapid development of low-altitude UAV remote sensing technology, UAV forestry remote sensing technology has become a new application technology in the new era. In the process of forest resources investigation, the response speed of UAV remote sensing processing system is very fast, and the collected digital elevation model (DEM) and mosaic data can be processed effectively in just a few minutes. At the same time, the spatial resolution of remote sensing images is higher than that of traditional satellite remote sensing images, it has been widely used in ecological data acquisition, vegetation mapping, biodiversity investigation and precision forestry.

In the ecosystem of arid areas, plants grow sparsely and individuals are small, so even high-resolution satellite remote sensing data can not fully meet the needs of accurate extraction of plant distribution, while UAV remote sensing technology can better solve the problem of resolution limitation. Qiu Yanning et al took the ecological restoration project area of Shapotou Grass in Zhongwei, Ningxia as the object, and used the Dajiang Mavic pro four-rotor UAV team to shoot orthophoto images (visible band 400m 700 nm) at the relative height of 30m, and obtained the orthophoto images of 2cm spatial resolution in the study area.

In the high-resolution images obtained, there are obvious differences among living plants, dry wheat straw and sandy land in the visible band, and the vegetation classification accuracy of each sample is more than 95%. Zhang Heyu obtained 1.5cm high resolution digital orthophoto image and digital elevation model (DEM) through UAV platform and motion structure modeling technology, and used decision tree algorithm to automatically estimate vegetation coverage based on orthophoto image, and superimposed with DEM data to analyze the spatial distribution characteristics of vegetation in Gobi area and its relationship with topography.

Under the background of the information age, the use of UAV forestry remote sensing advanced technology can achieve accurate resource survey in small areas. With the further development of science and technology, UAV forestry remote sensing technology will become more and more advanced and mature, which is of great significance to the future forestry development.

2.2 Forest information extraction

Forest parameter is an important index to reflect forest quality and management status. In the past, the acquisition of forest information was mainly carried out through the survey of sample plots. which had the advantages of low efficiency, great difficulty and poor accuracy, and was not conducive to large-scale investigation and research. As a new data acquisition method, forestry remote sensing technology can accurately and effectively extract many kinds of forest information, including forest tree species identification, forest volume estimation, forest canopy density estimation, foliar index, tree height and crown width, forest stand number and other information acquisition.

In recent years, forestry remote sensing technology plays an important role in the identification of forest tree species, especially in the classification of tree species under complex terrain with high precision. With the development of science and technology and the increase of data, the accuracy of tree species identification and classification by forestry remote sensing technology is getting higher and higher. Among them, Li Xusheng and others identified tree species in the subtropical natural secondary forest, and the overall identification accuracy reached 87.51%; Zhao Lin and others classified and identified the subtropical forest tree species, with an overall accuracy of 95.89%; Liu Lijuan and others identified the tree species in the nature reserve with an overall accuracy of 83.88%; Zhang Canghao and others identified the bamboo standing degree with the highest accuracy of 91.12%.

Although remote sensing technology has been effectively applied in the identification and classification of forest tree species, there are still few studies on forest ecosystems with rich tree species, high coverage and complex topography. For this reason, researchers use multi-source combination of data and optimization calculation methods to solve the difficulties of obtaining single data and difficult to extract information. At present, the application of forestry remote sensing technology in forest leaf area index inversion has been relatively mature, mainly using empirical model method and physical model method for remote sensing estimation of LAI.

Some scholars have successfully completed the estimation of vegetation leaf area index by using hyperspectral data under the condition of high model accuracy (R2 = 0.77). Some scholars have also implemented LAI monitoring by introducing machine learning model to improve the accuracy of LAI estimation. For example, some scholars apply BRDF correction image to the modeling and inversion of leaf area index, and the overall accuracy can reach 93.00%. Remote sensing technology has also been effectively applied in the extraction of forest crown width information and the calculation of stand volume. The DBH of trees is obtained by crown width-DBH correlation model, and then the method of calculating stand volume can effectively meet the accuracy requirements of forest resources survey.

2.3 Forest dynamic monitoring

Forest is a huge and complex ecosystem, which will continue to grow and multiply over time. Timely access to forest dynamic information is conducive to the effective management, protection and utilization of forest resources, and to deal with the emergency problems in the forest in time. Forestry remote sensing technology realizes all-weather real-time monitoring of forest dynamics.

It has been effectively applied in forest fire control, forest diseases and insect pests monitoring, forest desertification monitoring, forest wetland monitoring, forest rare wildlife resources monitoring and forest nutrient elements monitoring. In the aspect of forest fire prevention and control, through remote sensing technology, we can timely and accurately find the fire situation, obtain the fire point location information, detect the fire dynamics, calculate the fire area, and find out the nearest water source and road to the fire point to provide data support for the formulation of fire fighting tactics for fire officers and soldiers. Taking the "4 ·20" forest fire in Mianning as an example, remote sensing technology played an important role in the fire fighting.

Various departments quickly extracted forest fire elements through high-score remote sensing texture features, and obtained important information such as the location of water sources, rescue forces and rescue routes near the fire site, which effectively helped the emergency rescue of the fire and the prediction of hidden dangers. In the aspect of prevention and control of forest diseases and insect pests, Wu Hongkang et al used optical space remote sensing technology to monitor the pine caterpillar disaster in the 67000 square meter Masson pine stand in Guangxi, obtained the forest data at the flight altitude of 800m, and after the conversion between GIS and GPS data processing platform, adopted GPS navigation to arrive at the disaster area, and confirmed that 91.4% of the change in Masson pine stand quality was caused by pine caterpillar damage.

The misjudgment rate is only 1.9%. In other aspects of forest dynamic monitoring, Jindian et al used forestry remote sensing technology to establish monitoring and evaluation systems for different types of ecological restoration behaviors in 18 counties in the southern hilly and mountainous areas of Ganzhou, Jiangxi Province. Using the latest Landsat8 satellite images as the data source, Qin Zhangdan and others carried out the investigation and study of rocky desertification in karst areas by means of "the combination of indoor and field, human-computer interactive interpretation and computer automatic extraction".

2.4 Forestry law enforcement supervision

In forest resources management, there are often phenomena such as deforestation, indiscriminate felling and illegal occupation of woodland, while manual inspection not only has a large workload, but also has low accuracy and poor effectiveness. Forestry remote sensing technology can achieve real-time monitoring of the destruction of forest land, timely automatic alarm and effective evidence collection, directly locate the specific location of the incident through images, and accurately calculate the damaged forest land area, trees and other information. Taking timely law enforcement actions and remedial measures.

Forestry remote sensing technology has also played an important role in natural resources supervision. In 2020, the Natural Resources Department of Hunan Province adopted remote sensing technology in order to improve the efficiency of law enforcement, supervision and optimize the technical mode of natural resources supervision. Through remote sensing image data of different resolutions and different times during the year, we can achieve the requirements of early detection, early warning and early disposal of natural resources law enforcement supervision.

3 Conclusion

As a space exploration technology, remote sensing technology plays an active role in forest resources investigation, forest information extraction, forest dynamic monitoring and forestry law enforcement supervision. With the progress of science and technology, spectral information imaging, multi-directional optical detection, radar imaging multi-polarization, intelligent geoscience analysis, dynamic environmental research and quantitative resource research, it greatly improves the real-time, effectiveness and operation of forestry remote sensing technology, and makes it develop to the goal of three-dimensional, multi-level, multi-scale, multi-frequency, all-weather, high precision and high efficiency. With the rapid development of UAV technology, UAV remote sensing technology brings new methods and ideas to forestry work, which greatly improves the efficiency and quality of forestry work. In the future, in the research of UAV remote sensing technology, we can further expand the breadth and depth of the application of UAV in forestry, provide more powerful technical support for China's modern forestry, and make it better and more accurately serve the forestry production management.