Debris flow image recognition method

By setting early warning lines and sampling lines in the mudslide channel and extracting RGB values ​​using image processing technology, the problem of difficulty in monitoring the width and flow rate of mudslide in the prior art is solved, and accurate early warning and prevention and control effects are achieved.

CN119048972APending Publication Date: 2024-11-29INST OF MOUNTAIN HAZARDS & ENVIRONMENT CHINESE ACADEMY OF SCI
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410950426.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

It is difficult for the prior art to effectively monitor and early warning of the occurrence of mudslide flows, especially when the color changes and flow characteristics of mudslide flows in the channel are complex, and it is difficult to accurately judge the width and flow velocity of mudslide flows.

Method used

By setting up multiple warning lines and sampling lines in the mudslide channel, the RGB value of pixel points is extracted using image processing technology, and the color difference changes are calculated to judge the occurrence and flow rate of mudslide, so as to achieve early warning and prevention.

Benefits of technology

The accurate width and flow rate calculation of the mudslide flow is achieved, the efficiency of early warning and prevention of mudslide flow is improved, and the occurrence and scale of mudslide flow can be judged in a timely manner.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048972A_ABST
    Figure CN119048972A_ABST
Patent Text Reader

Abstract

The invention discloses a debris flow image recognition method, which comprises the following steps of: 1, continuously shooting a debris flow channel in an observation area by using a camera or a video camera to obtain a plurality of pictures or videos, and transmitting the pictures or videos to a computer for processing through a network to obtain a plurality of continuous pictures or video continuous frame screenshot pictures; and step 2, in a debris flow channel detection area in the picture, setting three or more early warning lines perpendicular to the flow direction of debris flow and a sampling line at the center of a channel, and setting a color difference early warning value of the early warning lines and the like. According to the debris flow image recognition method, based on the image color change characteristics of the debris flow channel at ordinary times and when the debris flow occurs, the flow velocity of the debris flow is calculated by processing the image, and the purposes of early warning and prevention of the debris flow are achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of debris flow observation in a channel, and more specifically to a debris flow image recognition method. Background Art

[0002] Debris flow often has the characteristics of sudden outbreak, fierceness and rapidity. It also has the dual effects of collapse, landslide and flood damage, and its degree of harm is more extensive and serious than the harm of single collapse, landslide and flood. When a debris flow occurs, prototype observation can be carried out in very few cases. An important way to study and prevent debris flow is to conduct prototype observation of debris flow. Fix a fixed-focus camera or video camera on the edge of the debris flow ditch or on a drone to shoot, analyze the image information of the debris flow movement to obtain the parameters required for debris flow prototype observation, and accumulate scientific data to play a positive role in promoting the development of debris flow discipline and the development of prevention and control work.

[0003] The existing channel debris flow has the following characteristics: First, when there is no debris flow in the debris flow ditch, there are stones of various sizes, mud, flowing water, and even plants in the ditch; second, when the debris flow occurs, due to the loss along the way encountered in the movement of the debris flow, the burst flow becomes thinner and the flow rate slows down. During the burst flow of debris flow, one burst advances farther than the other, and this process is called the "bed-laying process"; third, the movement form of the debris flow refers to the mud position process shape of the debris flow. The burst flow is the main movement form of viscous debris flow, and it has obvious head, body and tail. Therefore, most people call the above three parts "dragon head", "dragon body" and "dragon tail". The form of the burst flow: the front is high and steep, almost perpendicular to the riverbed; the head is like a sea tide with rolling waves and splashing waves, and gradually descends gently to the body and tail, and the entire flow surface is smooth and smooth.

[0004] Based on the above-mentioned characteristics of channel debris flow, this application designs a debris flow image recognition method. Summary of the invention

[0005] The purpose of the present invention is to provide a debris flow image recognition method, which uses the image color change characteristics of the debris flow channel during normal times and when the debris flow occurs as a basis, processes the image, and then calculates the width and flow rate of the debris flow, thereby achieving the purpose of debris flow early warning and prevention.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows: A debris flow image recognition method comprises the following steps: Step 1: Use a camera or video camera to continuously photograph the debris flow channel in the observation area to obtain a number of pictures or videos, and transfer the pictures or videos to a computer through a network for processing to obtain a number of continuous pictures or video continuous frame screenshots; Step 2: In the debris flow channel detection area in the image, set three or more warning lines perpendicular to the debris flow direction and a sampling line in the center of the channel, and set the warning line color difference warning value; Step 3: Extract the RGB values ​​of all pixels on the warning line. When the difference between the maximum and minimum values ​​of all RGB values ​​on the warning line is less than the warning value, a warning is issued. When the color difference warning values ​​of three or more sampling lines are simultaneously warned, it can be determined that a debris flow has occurred. Step 4. After determining that a debris flow has occurred in step 3, read the RGB values ​​of all pixels on the sampling line, divide the number of consecutive adjacent pixels with the same color difference on the warning line by the total number of pixels on the warning line, and multiply by the channel width to calculate the width of the debris flow when it passes the warning line. Then, determine the size of the faucet and its position in the channel by the size of the peak, continue to extract the RGB value peak position on the sampling line of the next image, and compare it with the position in the previous image to calculate the speed of the faucet.

[0007] As a further optimization of this solution, the camera or video camera is set on the left bank, right bank or in the air above the debris flow channel through a bracket or a drone.

[0008] As a further optimization of this solution, the sampling line intersects all the warning lines vertically, the distance between two adjacent pixels on the warning line is equal, the distance between two adjacent pixels on the sampling line that are far from the camera gradually increases, and conversely, the distance between two adjacent pixels that are close to the camera gradually decreases.

[0009] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention sets a number of warning lines and sampling lines based on the color change characteristics of images of debris flow channels during normal times and when debris flow occurs. By processing the images, it can be used to determine whether a debris flow occurs in the channel area, and then calculate the width and flow rate of the debris flow, thereby achieving the purpose of debris flow warning and prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a schematic diagram of the camera setting position structure of the present invention; Figure 2 This is a schematic diagram of the positions of the sampling line and the warning line of the present invention in the picture; Figure 3 A schematic diagram of extracting RGB values ​​of pixels at a certain cross section of the present invention; Figure 4 It is a schematic diagram of extracting RGB values ​​of pixel points in a certain longitudinal section of the present invention. DETAILED DESCRIPTION

[0011] In order to make the technical means, creative features, objectives and effects of the invention easy to understand, the present invention is further explained below with reference to specific diagrams.

[0012] Step 1: Figure 1 As shown, a camera position A or a camera position B is set on the downstream bank side of the debris flow channel of the debris flow observation station in the XX area, and the observed section is photographed by the camera to obtain several video continuous frame screenshots. Here, the debris flow channel is photographed by the camera, and the subsequent color extraction in the channel will find that the color changes greatly; Step 2: Use RTK to measure the camera coordinate point X, the two ends of the warning line Y11, Y12, Y21, Y22, Y31, Y32, and multiple points C1-Cn (n represents the number of sampling points) at equal intervals on the sampling line; at the same time, the camera records the image to determine the pixel point corresponding to each RTK measurement point, that is, the coordinates of each pixel point of the warning line and the coordinates of the RTK measurement point of the sampling line; Specifically, Figure 2 As shown, the warning line Y11 is set as the line segment between the pixel point (430, 420) and the pixel point Y12 (530, 420) on the picture; the warning line Y21 is set as the line segment between the pixel point (430, 490) and the pixel point Y22 (630, 490); the warning line Y31 is set as the line segment between the pixel point (430, 600) and the pixel point Y32 (830, 600); the sampling line is set as the line segment between the pixel point C1 (475, 410) and the pixel point C2 (475, 610), and the warning value is set to 30; Step 3: Read the normal and bed-making images, extract the RGB value of the warning line, and Figure 3 It can be analyzed that the color difference of the ditch bed in the Y420 cross section is very large at ordinary times; when the Y420 cross section is laid with debris flow, the color of the ditch bed is very uniform, and the color difference is very small. This is because the movement of the debris flow makes the water, mud, and stones in it fully mixed and the color is very uniform. When the difference between the maximum and minimum values ​​of all RGB on the warning line is less than the set warning value, an alarm is issued. When the color difference warning values ​​of three or more sampling lines are issued at the same time, it can be judged that a debris flow has occurred. Step 4: After determining that a debris flow has occurred, the RBG values ​​of all pixels on the sampling line are read continuously. Using the equal spacing between two adjacent pixels on the warning line, the number of consecutive adjacent pixels with the same color difference on the warning line is determined, divided by the total number of pixels on the warning line, and then multiplied by the channel width, that is, the width of the debris flow when it passes the warning line. Figure 4 From a certain X475 longitudinal section, it can be analyzed that the color of the debris flow ditch bed is uniform after the bed is laid. When the debris flow head passes by, the color difference changes and the sampling line has a peak. This is due to the strong turbulence of the head (the color difference becomes larger), which quickly transitions to the weak turbulence of the dragon body (the color difference becomes smaller) and the smooth laminar flow surface of the dragon tail (the color is uniform). Figure 4For the faucet of the debris flow, you will find that due to the steep front, strong turbulence and large color difference of the ditch bed after the bed is laid in front, when the debris flow faucet passes (recording time t1), a peak will appear in the RGB value. The size of the peak can be used to judge the scale of the faucet and its position in the ditch. Continue to extract the next image (recording time t2). The RGB value peak on the sampling line will move with the faucet. By comparing the position in the previous image, the speed of the faucet can be calculated, V = (the distance between sampling lines C2 and C1) / (t2-t1).

[0013] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A debris flow image recognition method, characterized in that: The steps include: Step 1: Use a camera or video camera to continuously photograph the debris flow channel in the observation area to obtain a number of pictures or videos, and transfer the pictures or videos to a computer through a network for processing to obtain a number of continuous pictures or video continuous frame screenshots; Step 2: In the debris flow channel detection area in the image, set three or more warning lines perpendicular to the debris flow direction and a sampling line in the center of the channel, and set the warning line color difference warning value; Step 3: Extract the RGB values ​​of all pixels on the warning line. When the difference between the maximum and minimum values ​​of all RGB values ​​on the warning line is less than the warning value, a warning is issued. When the color difference warning values ​​of three or more sampling lines are simultaneously warned, it can be determined that a debris flow has occurred. Step 4. After determining that a debris flow has occurred in step 3, read the RGB values ​​of all pixels on the sampling line, divide the number of consecutive adjacent pixels with the same color difference on the warning line by the total number of pixels on the warning line, and multiply by the channel width to calculate the width of the debris flow when it passes the warning line. Then, determine the size of the faucet and its position in the channel by the size of the peak, continue to extract the RGB value peak position on the sampling line of the next image, and compare it with the position in the previous image to calculate the speed of the faucet.

2. A debris flow image recognition method according to claim 1, characterized in that: The camera or video camera is arranged on the left bank, right bank or in the air above the debris flow channel through a bracket or a drone.

3. A debris flow image recognition method according to claim 2, characterized in that: The sampling line intersects all the warning lines perpendicularly, the distance between two adjacent pixels on the warning line is equal, the distance between two adjacent pixels on the sampling line far from the camera gradually increases, and conversely, the distance between two adjacent pixels close to the camera gradually decreases.

Citation Information

Cited By

  • Debris flow rapid quantitative identification method and application

    CN121919618A