Method for extracting and identifying laver culture empty raft frame

Abstract

The invention discloses a method for extracting and identifying a laver culture empty raft frame. The method is characterized by comprising the following steps: S1, acquiring original photos of a laver culture area and a raft frame; s2, image processing, including original photo quality inspection and orthoimage production; s3, performing neighborhood analysis calculation on the orthoimage preprocessed in the S2; s4, performing image threshold segmentation according to the information of the histogram of the neighborhood analysis calculation result; s5, grid-vector conversion is carried out on an image threshold segmentation result; s6, extracting a central point of a planar vector, and calculating the distance between a single element of a point-shaped vector and a peripheral element; s7, performing preliminary screening on the noise points by using a distance threshold value, and performing clustering analysis on the vector points after preliminary screening to form a clustering analysis result; s8, performing simplified edge processing on a clustering analysis result, and finally identifying and extracting the range and area of the laver culture empty raft frame; according to the method, the empty raft frame without laver growth is accurately recognized and extracted based on the unmanned aerial vehicle orthoimage, the problem that the empty raft frame without laver growth cannot be recognized through image index analysis and other methods is solved, and the blank of existing culture sea management work is filled.

Description

Technical Field

[0001] The invention belongs to the technical field of remote sensing, and in particular relates to a method for extracting and identifying an empty raft frame for laver cultivation. Background Art

[0002] In 2021, Lianyungang City was awarded the title of "China's Nori Capital" by the China Fisheries Association. It is the largest nori producing area in the country. At present, there are phenomena such as over-area farming, over-density farming, and unlicensed farming in the nori farming process, which have varying degrees of impact on sea area management, marine ecological environment, maritime transportation and emergency rescue, and increasing aquaculture production. To a certain extent, it restricts the high-quality development of nori farming.

[0003] The laver farming in Lianyungang City is characterized by a large area and a large number of rafts. At the same time, affected by factors such as manual harvesting, pests and diseases, and water quality, there are empty rafts in the laver farming rafts. Because they also occupy sea areas and are also subject to sea area management, it is difficult to obtain complete and accurate information through manual on-site field investigations. It is also difficult to quickly and accurately grasp the distribution and range of empty rafts from a macro perspective, and thus achieve accurate area measurement.

[0004] Monitoring using satellite remote sensing images and other technologies is easily affected by spatial resolution and limited by meteorological conditions, and cannot be effectively monitored. The inventors have searched technical personnel in this field to improve the accuracy of satellite remote sensing images and have developed various extraction methods, but have been unable to obtain a reasonable solution to the image index, so the extraction accuracy and type do not achieve the expected results at all.

[0005] The more common measurement method currently uses drones, but the photos obtained by drones cannot fully show the scope of the empty rafts for laver farming. The boundaries are blurred and difficult to determine, and there are even continuous areas. Further extraction and identification are needed to accurately determine the scope of the empty rafts for laver farming. Summary of the invention

[0006] The purpose of the invention is to design a method for extracting and identifying empty rafts for laver cultivation, which can quickly and accurately determine the distribution and range of the empty rafts and achieve effective monitoring.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] A method for extracting and identifying empty raft frames for laver cultivation, characterized in that the method is as follows:

[0009] S1: Get the original photos of the laver cultivation area and rafts;

[0010] S2: Image processing, including original photo quality inspection and orthophoto production;

[0011] S3: neighborhood analysis calculation is performed on the orthophotos preprocessed in S2;

[0012] S4: performing image threshold segmentation according to the information of the histogram of the neighborhood analysis calculation result;

[0013] S5: Perform raster-to-vector conversion on the result of image threshold segmentation;

[0014] S6: extract the center point of the surface vector and calculate the distance between a single element of the point vector and the surrounding elements;

[0015] S7: using the distance threshold to preliminarily screen the noise points, and performing cluster analysis on the preliminarily screened vector points to form a cluster analysis result;

[0016] S8: Simplify the edge processing of the cluster analysis results, and finally identify and extract the range and area of ​​the empty rafts for laver farming.

[0017] Furthermore, drone aerial photography is used to obtain original photos, and the drone specifically uses a Pegasus fixed-wing drone.

[0018] Furthermore, modeling was performed through the ArcGIS software Moddler tool to achieve image preprocessing.

[0019] Furthermore, the distance threshold represents the distance between adjacent points. If the distance between adjacent points is less than 2.5-3m, noise points exceeding 2.5-3m are judged to be noise points and can be removed.

[0020] Furthermore, the cluster analysis of the vector points is performed through distance observation, and the distance is less than 14-15m.

[0021] Through the above method, the following beneficial effects can be obtained:

[0022] This method accurately identifies and extracts empty raft frames without laver growth based on drone orthophotos, which solves the problem that empty raft frames without laver growth cannot be identified and extracted by other methods of non-site investigation. It fills the gap in the existing management of sea areas for aquaculture and is crucial to grasp the current status of laver farming and realize "refined and precise" management of sea areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is an orthophoto image taken by a drone.

[0024] Figure 2 It is a modeling flow chart.

[0025] Figure 3 This is the model calculation result diagram.

[0026] Figure 4 is the resulting image histogram.

[0027] Figure 5 It is the reclassification (image threshold segmentation) result map.

[0028] Figure 6 It is a cluster analysis model diagram.

[0029] Figure 7 This is the result diagram of the center point extraction of the surface vector.

[0030] Figure 8 It is a local enlarged image extracted from the center point of the surface vector.

[0031] Fig. 9 It is a distance table between vector points.

[0032] Fig.10 It is a schematic diagram of vector points after noise is removed.

[0033] Fig.11 This is the extraction result diagram of empty laver culture raft after cluster analysis.

[0034] Fig.12 This is the effect diagram after the extraction result is modified.

[0035] Fig.13 It is a method flow chart. DETAILED DESCRIPTION

[0036] The following embodiments are combined with the attached Figure 1-13 , the present invention is further described:

[0037] like Figure 1-13 As shown, a method for extracting and identifying empty raft frames for laver cultivation is characterized in that: the method is as follows:

[0038] S1: Get the original photos of the laver cultivation area and rafts;

[0039] S2: Image processing, including original photo quality inspection and orthophoto production;

[0040] S3: neighborhood analysis calculation is performed on the orthophotos preprocessed in S2;

[0041] S4: performing image threshold segmentation according to the information of the histogram of the neighborhood analysis calculation result;

[0042] S5: converting the result of image threshold segmentation from raster to vector;

[0043] S6: extract the center point of the surface vector and calculate the distance between a single element of the point vector and the surrounding elements;

[0044] S7: using the distance threshold to preliminarily screen the noise points, and performing cluster analysis on the preliminarily screened vector points to form a cluster analysis result;

[0045] S8: Simplify the edge processing of the cluster analysis results, and finally identify and extract the range and area of ​​the empty rafts for laver farming.

[0046] Furthermore, drone aerial photography is used to obtain original photos, and the drone specifically uses a Pegasus fixed-wing drone.

[0047] Furthermore, modeling was performed through the ArcGIS software Moddler tool to achieve image preprocessing.

[0048] Furthermore, the distance threshold represents the distance between adjacent points. If the distance between adjacent points is less than 3m, noise points exceeding 3m are judged to be noise points and can be removed.

[0049] Furthermore, the cluster analysis of the vector points is performed through distance observation, and the distance is less than 15m.

[0050] Example:

[0051] Fig.13 The figure shows a method for extracting and identifying the empty raft frame of laver cultivation. Taking the laver cultivation area in Lianyungang City as an example, the specific steps of the method are as follows:

[0052] S1: Aerial photography to obtain original drone photos of laver cultivation areas; combined with the unit's regular aquaculture sea supervision, use Pegasus F300 / Pegasus V300 fixed-wing drones, and use the aircraft's own process software, combined with the aquaculture season, to reasonably plan flight routes and carry out drone aerial photography of laver cultivation areas;

[0053] S2: UAV image processing, including quality inspection of original UAV photos and production of UAV orthophotos; using the aircraft's own process data processing software, quality inspection of original photos and production of orthophotos are carried out to obtain Figure 1 Drone orthophotos shown;

[0054] S3: Modeling is performed through the ArcGIS software Moddler tool to perform preprocessing work such as original photo inspection and aerial photography quality inspection of the image. The preprocessed orthophotos are then subjected to neighborhood analysis and calculation. Figure 2 Demonstrate the modeling process. Figure 3 Get the model calculation result diagram;

[0055] S4: Histogram of the results calculated based on neighborhood analysis ( Figure 4 The information of the image is reclassified (image threshold segmentation) to distinguish the raft poles of the empty raft from the sea water from the perspective of imaging, which is convenient for subsequent identification, extraction and measurement. The reclassification results are detailed in Figure 5 As shown;

[0056] S5: Perform grid-to-vector conversion (referred to as grid-to-vector conversion) on the reclassification results and convert them into vector data. In the reclassification results, the raft poles of the empty rafts can be seen, that is, spots (surfaces) in the circle. This solves the problem of large number of images and cumbersome data calculation. At the same time, through vector processing, the subsequent accurate calculation of the raft area and accurate identification of the number of rafts can be achieved;

[0057] S6: Use the vector surface to point tool to extract the center point of the surface vector and perform analysis and calculation on the subsequent clustering analysis, that is, the point aggregation situation. For details, see Figure 7 and Figure 8 The center point extraction result of the mid-surface vector;

[0058] S7: From the results, it can be found that the rafts without laver culture show a point-like clustering phenomenon. Therefore, the next step is to calculate the shortest distance between all vector points in these images and extract the point "clustering" phenomenon by setting a threshold. For details, see Fig. 9 As shown;

[0059] S8: Use the distance threshold (distance between adjacent points is less than 2.5m) to preliminarily screen noise points, that is, all points whose distances from a certain point to surrounding elements exceed 2.5m can be considered as noise points and can be deleted, effectively eliminating the influence of noise points caused by seawater, environment, aircraft cameras, etc. on the recognition results, and improving the accuracy of raft recognition;

[0060] S9: Use the aggregate points tool in the ArcGIS toolbox to perform cluster analysis on the results after noise removal and cluster analysis on the vector points after initial screening (with a distance less than 15m) to form cluster analysis results;

[0061] S10: The cluster analysis results were modified using the simplified surface tool in the ArcGIS tool set to identify and extract the range and area of ​​the empty rafts for laver farming. Fig.12 shown.

[0062] The above are all preferred embodiments of the present invention. For ordinary technicians in this technical field, without departing from the principle of the present invention, various equivalent modifications to the present invention belong to the protection scope of the claims attached to this application.

Claims

1. A method for extracting and identifying empty rafts for laver cultivation, characterized in that: The method is as follows: S1: Get the original photos of the laver cultivation area and rafts; S2: Image processing, including original photo quality inspection and orthophoto production; S3: neighborhood analysis calculation is performed on the orthophotos preprocessed in S2; S4: performing image threshold segmentation according to the information of the histogram of the neighborhood analysis calculation result; S5: Perform raster-to-vector conversion on the result of image threshold segmentation; S6: extract the center point of the surface vector and calculate the distance between a single element of the point vector and the surrounding elements; S7: using the distance threshold to preliminarily screen the noise points, and performing cluster analysis on the preliminarily screened vector points to form a cluster analysis result; S8: Simplify the edge processing of the cluster analysis results, and finally identify and extract the range and area of ​​the empty rafts for laver farming.

2. The method for extracting and identifying an empty raft frame for laver cultivation according to claim 1, characterized in that: The original photos are obtained by aerial photography using a drone, and the drone used is a Pegasus fixed-wing drone.

3. The method for extracting and identifying an empty raft frame for laver cultivation according to claim 1, characterized in that: Modeling and image preprocessing are performed using the ArcGIS software Moddler tool.

4. The method for extracting and identifying an empty raft frame for laver cultivation according to claim 1, characterized in that: The distance threshold represents the distance between adjacent points. If the distance between adjacent points is less than 2.5-3m, noise points exceeding 2.5-3m are judged as noise points and can be removed.

5. The method for extracting and identifying an empty raft frame for laver cultivation according to claim 1 or 4, characterized in that: The cluster analysis of the vector points is performed by distance observation, and the distance is less than 10-15m.

Images