CN-116778313-B - Method and system for identifying field road

CN116778313BCN 116778313 BCN116778313 BCN 116778313BCN-116778313-B

Abstract

The invention relates to a field road identification method and a field road identification system, wherein the method comprises the steps of obtaining field image information by using an unmanned aerial vehicle; determining a segmentation threshold value by using a golden section method according to the field image information, segmenting by using the segmentation threshold value according to the field image information to obtain segmented roads, and determining the field roads by using an accumulation model of mathematical morphology according to the segmented roads. The method can quickly realize automatic identification of farmland road information.

Inventors

HU YUEMING
CHEN LIANCHENG
XU TIAN
ZHANG FEIYANG
XIAO JIAMING

Assignees

广州市华南自然资源科学技术研究院

Dates

Publication Date: 20260505
Application Date: 20220308

Claims (8)

1. A method for identifying a road in a field, comprising: Acquiring field image information by using an unmanned aerial vehicle; Determining a segmentation threshold value by using a golden section method according to the field image information; dividing the field image information by utilizing the dividing threshold value to obtain a divided road; The method for determining the field road by using the accumulation model of the mathematical morphology according to the segmented road specifically comprises the steps of determining a gray value accumulation model according to continuous pixel points of the segmented road, wherein the continuous pixel points comprise continuous pixel points along the row direction and continuous pixel points along the column direction, and determining the field road according to the gray value accumulation model.
2. The field road identification method according to claim 1, further comprising, after the acquiring of the field image information with the unmanned aerial vehicle: carrying out gray scale processing on the field image information to obtain a gray scale image; and carrying out feature analysis on the gray level map to obtain a gray level value range.
3. The method for identifying a road in a field according to claim 2, wherein determining the segmentation threshold by golden section method based on the field image information comprises: and determining a segmentation threshold value by using a golden section method according to the gray value range of the field image information.
4. The field road identification method according to claim 1, further comprising, after the dividing by the dividing threshold according to the field image information, the steps of: And carrying out binarization processing on the segmented road.
5. A field road identification system, comprising: The acquisition module is used for acquiring field image information by using the unmanned aerial vehicle; the segmentation threshold determining module is used for determining a segmentation threshold by using a golden section method according to the field image information; the segmentation module is used for segmenting by utilizing the segmentation threshold according to the field image information to obtain segmented roads; the field road determining module is used for determining a field road according to the segmented road by using an accumulation model of mathematical morphology; the field road determining module specifically comprises: the system comprises a gray value accumulation model determining unit, a gray value accumulation model determining unit and a gray value accumulation model determining unit, wherein the gray value accumulation model determining unit is used for determining a gray value accumulation model according to continuous pixel points of the segmented road, and the continuous pixel points comprise continuous pixel points along the row direction and continuous pixel points along the column direction; And the field road determining unit is used for determining the field road according to the gray value accumulation model.
6. The field road identification system of claim 5, further comprising: the gray processing module is used for carrying out gray processing on the field image information to obtain a gray image; And the characteristic analysis module is used for carrying out characteristic analysis on the gray level image to obtain a gray level value range.
7. The field road identification system of claim 6, wherein the segmentation threshold determination module specifically comprises: and the segmentation threshold determining unit is used for determining a segmentation threshold by using a golden section method according to the gray value range of the field image information.
8. The field road identification system of claim 5, further comprising: And the binarization processing module is used for carrying out binarization processing on the segmented road.

Description

Method and system for identifying field road Technical Field The invention relates to the field of image recognition, in particular to a method and a system for recognizing a field road. Background Basic farmland construction importance. The national importance of basic farmland construction, the cost of huge capital for high-standard farmland construction. Monitoring the quality of farmland infrastructure is a key means of high-standard farmland construction, and the high-standard farmland construction of large-scale agriculture can be effectively monitored by adopting an informatization means. And farmland information automatic acquisition is performed by using the unmanned aerial vehicle, so that the acquisition and recognition efficiency is improved. The accessibility of farmland roads is one of the important standards of high-standard farmland, and road information must be automatically collected to automatically calculate the accessibility of farmland roads. Unmanned aerial vehicle remote sensing plays an increasingly important role in various fields of national economy and social development such as agriculture, ecological environment, new rural construction planning, natural disaster monitoring, safety consolidation, water yield, mineral resource exploration and the like by virtue of the advantages of full-time, real-time, high resolution, flexible maneuvering, high cost performance and the like, and becomes an emerging development direction following a satellite remote sensing technology. The efficiency of farmland road information collection discernment has been improved greatly with unmanned aerial vehicle automatic acquisition farmland road information. The automatic collection and identification of farmland road information has problems. The automatic extraction of road information and the automatic identification of roads in the current traffic field are popular researches in the current traffic field. Many research methods for road identification exist, but various imperfect and unresolved defects exist, so automatic road identification has been a high focus of attention for experts for a long time. The same problem is more true for farmland roads that have not received attention from the expert. At present, the extraction and identification of farmland road information are basically blank, so that the automatic acquisition of farmland road information and road identification have more research significance. The identification of the field roads cannot be directly applied to the existing urban road identification method. The application requirements and the road information characteristics of the field road information of the high-standard farmland are different from those of urban roads, so that the automatic identification of the field roads cannot be directly applied to the existing urban road identification method. The urban road identification application requirements are mainly unmanned, navigation, intelligent traffic, maps and the like, the farmland road application requirements are access conditions of roads and the like, the urban road application system can be used for fully configuring various high-grade equipment, the urban road can only carry the simplest equipment due to poor road monitoring environment conditions, the urban road can be arranged in a better network environment, the urban road is in the weak environment of the network, the urban road environment is relatively standard, the farmland is poor in environment, the characteristics of land features around the roads are different, traffic lines of the urban are different from the roads of the farmland in road color, shape and texture, and the system which is used for carrying out field acquisition on a farmland is more required to be used for fast acquisition and can carry out a field decision function due to wide farmland amplitude personnel, high material resources and time cost for round trip field acquisition for high-standard farmland construction project monitoring, large-scale farmland management and planning and the like. So the automatic extraction and identification of the road information of the research farmland is a new problem under the new situation of agricultural modernization. Currently, many conventional road identification methods are available, and the most common method is to search for corresponding roads from a google map. Road maps of various places can be searched by utilizing the google map. However, for new projects of farmland facilities, roads of the new projects cannot be immediately constructed on a google map, remote sensing images of moreover, satellites are only in transit for a limited number of times each year, and weather clouds have great influence on image quality during each transit, so that it is not practical to extract the information of the new roads of farmland infrastructure in real time by utilizing satellite remote sensing. The remote sensing or unmanned aerial vehicle im