Unmanned aerial vehicle water area image automatic splicing optimization algorithm

Abstract

The invention discloses an unmanned aerial vehicle water area image automatic splicing optimization algorithm in the technical field of black and odorous water body classification and recognition. Comprising the steps of establishing an image recognition system, performing recognition training on the image recognition system, setting a flight path of an unmanned aerial vehicle, presetting image acquisition time, performing image acquisition on black and odorous water in a specified area, recognizing acquired black and odorous water image data, obtaining recognition results and classifying the recognition results. According to the method, image acquisition is performed on the black and odorous water body in the specified area through the unmanned aerial vehicle, the acquired image data are transmitted to the identification system, the acquired black and odorous water body image data are identified, and then the identification result is acquired and classified, so that the black and odorous water body classification and identification technology based on the time sequence unmanned aerial vehicle image can be completed. The water body quality and the distribution condition of the black and odorous water body in the area are known, the black and odorous water body is conveniently, timely and effectively treated, and the ecological environment of the water body is effectively recovered.

Description

Technical Field

[0001] The present invention relates to the technical field of black and odorous water body classification and identification, and specifically to an automatic splicing optimization algorithm for drone water area images. Background Art

[0002] Black and odorous water bodies, formed due to excessive pollution, have lost their ecological function. Urban black and odorous water bodies are a water environment issue that residents strongly complain about. They not only damage the urban living environment but also seriously affect the city's image. In recent years, calls for "letting the mayor swim in the river" have reflected the public's strong desire to address and control urban black and odorous water bodies.

[0003] The classification and identification of black and odorous water bodies are mainly carried out through field surveys, which are a combination of human subjective judgment and water quality testing. However, field surveys in large areas consume a lot of manpower and material resources. Due to the wide range of water areas, the degree of black and odorous water bodies in different locations within the water area is different, and the black and odorous water bodies in the entire area are unevenly distributed. The classification and identification of black and odorous water bodies are not comprehensive enough, and it is inconvenient to treat black and odorous water bodies in a targeted manner. To this end, we propose an automatic stitching optimization algorithm for drone water area images. Summary of the Invention

[0004] The purpose of the present invention is to provide an automatic stitching optimization algorithm for drone water area images to solve the problems of uneven distribution of black and odorous water bodies and incomplete classification and identification raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an automatic splicing optimization algorithm for drone water imagery, comprising establishing an image recognition system, training the image recognition system, setting a drone flight path, presetting an image acquisition time, acquiring images of black and odorous water bodies within a specified area, identifying the acquired black and odorous water body image data, obtaining recognition results, and classifying the recognition results. The black and odorous water body classification and recognition technology based on time-series drone images includes the following steps:

[0006] Step 1: Establish an image recognition system: Establish an image recognition system based on the degree of black and odorous water bodies;

[0007] Step 2: Train the image recognition system: Collect a wide range of water body photos and train the image recognition system to enable it to recognize the collected black and odorous water images and output the recognition results;

[0008] Step 3: Set the drone flight path: According to satellite navigation positioning, set the drone's take-off point and the end point for black and odorous water body image collection, and set the drone's flight path and the image collection range of the black and odorous water body;

[0009] Step 4, preset image acquisition time: through the image acquisition module, preset the image acquisition time, so that the drone can capture images of black and odorous water bodies at the specified time during flight;

[0010] Step 5: Capture images of black and odorous water bodies within the specified area: The drone flies along a preset path based on satellite navigation positioning. During the flight, the drone captures images of the black and odorous water bodies in a time sequence of pre-set image capture times;

[0011] Step 6: Identify the collected black and odorous water body image data: Use the image recognition system to identify the collected black and odorous water body image and output the result;

[0012] Step 7, obtain the recognition results and classify the recognition results: According to step 6, obtain the output black and odorous water body image recognition results, classify them according to the output results, and fully understand the water quality and distribution of black and odorous water bodies in the entire area.

[0013] Preferably, in step 1, the black and odorous water body can be divided into two levels according to the degree of black and odor, namely, "mild black and odorous" and "severe black and odorous". The grading standards of "mild black and odorous" and "severe black and odorous" are as follows:

[0014] If the water transparency is between 25 cm and 10 cm, it is mildly black and smelly; if the water transparency is less than 10 cm, it is severely black and smelly.

[0015] Preferably, in step 2, water body images are widely collected as training samples, and the actual conditions of the training samples are used as a control group to perform recognition training on the image recognition system, the training results are output for classification statistics, the images with recognition errors are extracted, and the system is improved to reduce recognition errors.

[0016] Preferably, in step 3, the travel path of the drone is set according to satellite navigation positioning, and the range for image collection of black and odorous water bodies is set so that the drone operates within the specified area.

[0017] Preferably, in step 4, the time for capturing images of the black and odorous water body is preset through the image capture module, and the images of the black and odorous water body are automatically captured at the specified time, and the time interval between each two image captures is the same. The image capture module is a GoPro sports camera.

[0018] Preferably, in step 5, the drone flies from the take-off point to the end point of black and odorous water body image collection according to the flight path set in step 3, collects images of the black and odorous water body in chronological order according to the image collection time preset in step 2, and transmits the collected black and odorous water body image data to the recognition system.

[0019] Preferably, in step 6, the black and odorous water body image data transmitted by the image acquisition module is received, the black and odorous water body image collected by the drone is recognized through the set image recognition system, and the recognition result is output.

[0020] Preferably, in step 7, classification is performed based on the obtained black and odorous water image recognition results to understand the water quality and distribution of black and odorous water bodies in the area, so as to facilitate timely and effective treatment of black and odorous water bodies and effectively restore the water ecological environment.

[0021] Compared with the existing technology, the beneficial effects of the present invention are: using drones to collect images of black and odorous water bodies in a specified area, transmitting the collected image data to a recognition system, identifying the collected black and odorous water body image data and outputting the recognition results, then obtaining and classifying the recognition results, and judging the degree of black and odorous water bodies. The black and odorous water body classification and recognition technology based on time series drone images can be completed, thereby understanding the water quality and distribution of black and odorous water bodies in the area, conveniently and promptly and effectively treating black and odorous water bodies, and effectively restoring the water body ecological environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flow chart of an automatic stitching optimization algorithm for drone water imagery of the present invention. DETAILED DESCRIPTION

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0024] The present invention provides a technical solution: an automatic splicing optimization algorithm for drone water images, comprising establishing an image recognition system, performing recognition training on the image recognition system, setting a drone flight path, presetting an image acquisition time, acquiring images of black and odorous water bodies within a specified area, identifying the acquired black and odorous water body image data, obtaining recognition results, and classifying the recognition results. The black and odorous water body classification and recognition technology based on time series drone images includes the following steps:

[0025] Step 1: Establish an image recognition system: Establish an image recognition system based on the degree of black and odorous water bodies;

[0026] Step 2: Train the image recognition system: Collect a wide range of water body photos and train the image recognition system to enable it to recognize the collected black and odorous water images and output the recognition results;

[0027] Step 3: Set the drone flight path: According to satellite navigation positioning, set the drone's take-off point and the end point for black and odorous water body image collection, and set the drone's flight path and the image collection range of the black and odorous water body;

[0028] Step 4, preset image acquisition time: through the image acquisition module, preset the image acquisition time, so that the drone can capture images of black and odorous water bodies at the specified time during flight;

[0029] Step 5: Capture images of black and odorous water bodies within the specified area: The drone flies along a preset path based on satellite navigation positioning. During the flight, the drone captures images of the black and odorous water bodies in a time sequence of pre-set image capture times;

[0030] Step 6: Identify the collected black and odorous water body image data: Use the image recognition system to identify the collected black and odorous water body image and output the result;

[0031] Step 7, obtain the recognition results and classify the recognition results: According to step 6, obtain the output black and odorous water body image recognition results, classify them according to the output results, and fully understand the water quality and distribution of black and odorous water bodies in the entire area.

[0032] Furthermore, in step 1, the black and odorous water body can be divided into two levels according to the degree of black and odor: "mild black and odorous" and "severe black and odorous". The classification standards of "mild black and odorous" and "severe black and odorous" are as follows:

[0033] If the water transparency is between 25 cm and 10 cm, it is mildly black and smelly; if the water transparency is less than 10 cm, it is severely black and smelly.

[0034] Furthermore, in step 2, water body images are widely collected as training samples, and the actual conditions of the training samples are used as a control group to train the image recognition system. The training results are output for classification and statistics, the images with recognition errors are extracted, and the system is improved to reduce recognition errors.

[0035] Furthermore, in step 3, the travel path of the UAV is set according to the satellite navigation positioning, and the range for collecting images of the black and odorous water body is set so that the UAV operates within the specified area.

[0036] Furthermore, in step 4, the time for capturing images of the black and odorous water body is preset through the image capture module, and the images of the black and odorous water body are automatically captured at the specified time, and the time interval between each two image captures is the same. The image capture module is a GoPro sports camera.

[0037] Furthermore, in step 5, the UAV flies from the take-off point to the end point of black and odorous water body image collection according to the flight path set in step 3, collects images of the black and odorous water body in chronological order according to the image collection time preset in step 2, and transmits the collected black and odorous water body image data to the recognition system.

[0038] Furthermore, in step 6, the black and odorous water body image data transmitted by the image acquisition module is received, and the black and odorous water body image collected by the drone is recognized through the set image recognition system, and the recognition result is output.

[0039] Furthermore, in step 7, classification is performed based on the obtained black and odorous water image recognition results to understand the water quality and distribution of black and odorous water bodies in the area, so as to facilitate timely and effective treatment of black and odorous water bodies and effectively restore the water ecological environment.

[0040] An automatic stitching optimization algorithm for drone water area images adopted by the present invention achieves technical effects through the following steps: first, an image recognition system is established: an image recognition system is established according to the degree of black and odorous water bodies; second, recognition training is performed on the image recognition system: water body photos are widely collected, and recognition training is performed on the image recognition system, so that the recognition system can recognize the collected black and odorous water body images and output the recognition results; then, the drone flight path is set: according to satellite navigation positioning, the take-off point of the drone and the end point of the black and odorous water body image collection are set, and the flying path of the drone and the image collection range of the black and odorous water body are set; then, the image collection time is preset: through the image collection module, the image collection time is preset, so that the drone can collect images of black and odorous water bodies at the specified time when flying; then, the black and odorous water bodies in the specified area are collected: the drone flies along the preset path according to the satellite navigation positioning, and during the flight of the drone, the drone is The black and odorous water bodies are collected in the time sequence of image collection time; then, the collected black and odorous water body image data is identified: the collected black and odorous water body images are identified by the image recognition system and the results are output; finally, the recognition results are obtained and the recognition results are classified: according to step 6, the output black and odorous water body image recognition results are obtained, and the output results are classified to fully understand the water quality of the entire area and the distribution of black and odorous water bodies. The black and odorous water bodies in the specified area are collected by drones, and the collected image data are transmitted to the recognition system. The collected black and odorous water body image data are identified and the recognition results are output. Then, the recognition results are obtained and the recognition results are classified to determine the black and odorous degree of the black and odorous water bodies. The black and odorous water body classification and recognition technology based on time series drone images can be completed, thereby understanding the water quality and distribution of black and odorous water bodies in the area, conveniently and effectively treating black and odorous water bodies in a timely and effective manner, and effectively restoring the water body ecological environment.

[0041] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An automatic stitching optimization algorithm for drone water imagery, comprising establishing an image recognition system, training the image recognition system, setting a drone flight path, presetting an image acquisition time, acquiring images of black and odorous water bodies within a specified area, identifying the acquired black and odorous water body image data, obtaining recognition results, and classifying the recognition results, characterized by: The black and odorous water body classification and identification technology based on time series drone images includes the following steps: Step 1: Establish an image recognition system: Establish an image recognition system based on the degree of black and odorous water bodies; Step 2: Train the image recognition system: Collect a wide range of water body photos and train the image recognition system to enable it to recognize the collected black and odorous water images and output the recognition results; Step 3: Set the drone flight path: According to satellite navigation positioning, set the drone's take-off point and the end point for black and odorous water body image collection, and set the drone's flight path and the image collection range of the black and odorous water body; Step 4, preset image acquisition time: through the image acquisition module, preset the image acquisition time, so that the drone can capture images of black and odorous water bodies at the specified time during flight; Step 5: Capture images of black and odorous water bodies within the specified area: The drone flies along a preset path based on satellite navigation positioning. During the flight, the drone captures images of the black and odorous water bodies in a time sequence of pre-set image capture times; Step 6: Identify the collected black and odorous water body image data: Use the image recognition system to identify the collected black and odorous water body image and output the result; Step 7, obtain the recognition results and classify the recognition results: According to step 6, obtain the output black and odorous water body image recognition results, classify them according to the output results, and fully understand the water quality and distribution of black and odorous water bodies in the entire area.

2. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 1, black and odorous water bodies can be divided into two levels according to the degree of black and odor: "mild black and odorous" and "severe black and odorous". The classification standards for "mild black and odorous" and "severe black and odorous" are as follows: If the water transparency is between 25 cm and 10 cm, it is mildly black and smelly; if the water transparency is less than 10 cm, it is severely black and smelly.

3. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 2, water body images are widely collected as training samples, and the actual conditions of the training samples are used as a control group to train the image recognition system. The training results are output for classification and statistics, the images with recognition errors are extracted, and the system is improved to reduce recognition errors.

4. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 3, the drone's travel path is set according to satellite navigation positioning, and the range for image collection of black and odorous water bodies is set so that the drone operates within the specified area.

5. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 4, the time for collecting images of black and odorous water bodies is preset through the image collection module, and the images of black and odorous water bodies are automatically collected at the specified time, and the time interval between each two image collections is the same. The image collection module is a GoPro sports camera.

6. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 5, the UAV flies from the take-off point to the end point of black and odorous water body image collection according to the flight path set in step 3, collects images of the black and odorous water body in chronological order according to the image collection time preset in step 2, and transmits the collected black and odorous water body image data to the recognition system.

7. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 6, the black and odorous water body image data transmitted by the image acquisition module is received, and the black and odorous water body image collected by the drone is recognized through the set image recognition system, and the recognition result is output.

8. The automatic stitching optimization algorithm for drone water imagery according to claim 1 is characterized by: In step 7, the black and odorous water bodies are classified according to the obtained image recognition results to understand the water quality and distribution of black and odorous water bodies in the area, so as to facilitate timely and effective treatment of black and odorous water bodies and effectively restore the water ecological environment.

Images