A method for determining the boundary line between a ship and a water surface

Abstract

The present invention proposes a method for determining the boundary line between a ship and the water surface, comprising the following steps: applying a deep learning algorithm to pre-determine whether the current recognition scene belongs to a double interface or a triple interface, and the color composition of the main body of the hull; detecting the hull edge of the image by RGB and HSV to obtain a contour line object group, applying the hull color matrix to comprehensively determine the contour line of the boundary between the hull and the water surface, and preliminarily locating the water surface boundary line A; enhancing the contrast of the images on both sides of the interface and the contrast between the boundary line and the images on both sides; dividing the image area processed by HSV superposition and RGB superposition and longitudinally, and taking the water surface boundary line A as the low point, obtaining a slice group, measuring each slice separately, and calculating to obtain a value group; fitting the value group into a boundary line array with the same width as the output image and outputting it for water gauge measurement. The present invention can effectively solve the problem of accurate positioning of the water surface boundary line in ship water gauge measurement.

Description

Technical Field

[0001] The invention relates to the technical field of image recognition, in particular to a method for determining a boundary line between a ship and a water surface. Background Art

[0002] At present, bulk cargo delivery is settled by water gauge measurement. Water gauge measurement is based on the average value of six scales evenly distributed around the ship. In this way, the gross weight and tare weight of the ship are measured and calculated respectively, and the difference is the weight of the cargo. Although water gauge measurement is intuitive, it is often affected by factors such as river waves and human visual inspection, which often leads to large errors in measurement results and is easy to be unfair in the settlement process. In addition, the entire measurement process basically relies on manual measurement. On-site measurement and calculation generally take nearly 3 hours, and the low efficiency affects the unloading time.

[0003] Using computer image recognition technology to process video data shot by mobile phones to obtain water gauge readings can reduce the error of manual visual inspection. The water gauge measurement process includes four main processes: water gauge positioning, water surface boundary positioning, water gauge readings, and loading and unloading volume calculation. The water surface boundary line refers to the boundary between the ship and the water surface. The purpose of water surface boundary positioning is to determine the intersection point between the ship and the water surface, which is used to determine the water gauge reading.

[0004] In conventional methods, the water boundary line in the water gauge area of ​​the video image is often difficult to confirm. It is affected by the following factors: the interaction between the clarity of the water body and the light, water stains on hulls of different colors, the reflection of the hull on the water surface is similar to the hull, the hull's own welding lines affect the determination of the water boundary line, sunlight reflection, the presence of three interfaces between the shore, the water surface and the hull in some videos, and the upper and lower waterlines of some hulls are close to the water surface line. Summary of the invention

[0005] The purpose of the present invention is to provide a safe and reliable method for determining the boundary line between a ship and a water surface.

[0006] The technical solution adopted by the present invention to solve the above technical problems is: the process of identifying the boundary between the ship and the water surface is as follows:

[0007] Scene recognition->edge detection->HSV overlay processing->RGB overlay processing->boundary line measurement->boundary line fitting; the input of water surface boundary positioning is the shot video, water level area image, and water level object set B, and the output is the boundary line fitting curve.

[0008] Scene recognition: Use deep learning algorithms to pre-determine whether the current recognition scene belongs to a double interface or a triple interface, and the main color composition of the hull.

[0009] Edge detection: Detect the edge of the hull using the RGB and HSV of the image to obtain the contour line object group, use the hull color matrix to comprehensively judge the contour line of the boundary between the hull and the water surface, and preliminarily locate the water surface boundary line A

[0010] The HSV overlay processing and the RGB overlay processing are mainly used to enhance the contrast of the images on both sides of the interface and the contrast between the boundary line and the images on both sides.

[0011] Boundary line measurement: The image area that has undergone HSV overlay processing and RGB overlay processing is divided longitudinally, and the water surface boundary line A is used as the low point to obtain slice grouping. Each slice is measured separately to calculate the value group.

[0012] Boundary line fitting curve: Fit the value group into a boundary line array with the same width as the output image, which is used for water level measurement.

[0013] Compared with the prior art, the advantages of the present invention are as follows: the present invention solves the problem of accurate positioning of the water surface boundary line in ship water level measurement, the shooting distance can be arbitrarily adjusted within 1-20 meters, reaching the measurement error range of 5 thousandths, and the recognition error compared with human eye observation is <2cm, with high anti-interference ability and a recognition rate of more than 95%. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0015] The present invention is further described below with reference to the accompanying drawings and specific embodiments.

[0016] like Figure 1 As shown in the figure, the process of identifying the boundary between the ship and the water surface is as follows:

[0017] Scene recognition->edge detection->HSV overlay processing->RGB overlay processing->boundary line measurement->boundary line fitting; the input of water surface boundary positioning is the shot video, water level area image, and water level object set B, and the output is the boundary line fitting curve.

[0018] Scene recognition: A deep learning algorithm is used to pre-determine whether the current recognition scene belongs to a double interface or a triple interface, and the color composition of the hull body. This embodiment adopts a neural network learning algorithm. The neural network training parameters include the number of iterations: 60,000; the learning rate: 0.2; the weight reduction coefficient: 0.0003. The neural network training uses a feedforward algorithm for parameter learning.

[0019] Edge detection: Detect the edge of the hull using the RGB and HSV of the image to obtain the contour line object group, use the hull color matrix to comprehensively judge the contour line of the boundary between the hull and the water surface, and preliminarily locate the water surface boundary line A

[0020] The HSV overlay processing and the RGB overlay processing are mainly used to enhance the contrast of the images on both sides of the interface and the contrast between the boundary line and the images on both sides.

[0021] Boundary line measurement: The image area that has undergone HSV and RGB overlay processing is divided longitudinally, and the water surface boundary line A is used as the low point to obtain slice grouping. Each slice is measured separately to calculate the value group.

[0022] Boundary line fitting curve: Fit the value group into a boundary line array with the same width as the output image, which is used for water level measurement.

[0023] The present invention solves the problem of accurate positioning of the water surface boundary line in ship water level measurement. The shooting distance can be arbitrarily adjusted within 1-20 meters, achieving a measurement error range of 5 thousandths, and the recognition error compared with human eye observation is <2cm. It has high anti-interference ability and a recognition rate of more than 95%.

[0024] The contents not described in detail in the specification of the present invention belong to the prior art known to the professional and technical personnel in this field.

[0025] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0026] Although the 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 the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for determining the boundary between a ship and a water surface, characterized in that: The following steps are involved: Step 1, scene recognition: Apply deep learning algorithms to pre-determine whether the current recognition scene belongs to a double interface or a triple interface, and the main color composition of the hull; Step 2, edge detection: Detect the edge of the hull from the RGB and HSV of the image to obtain the contour line object group, use the hull color matrix to comprehensively judge the contour line of the boundary between the hull and the water surface, and preliminarily locate the water surface boundary line A; Step 3, HSV overlay processing and RGB overlay processing: enhance the contrast of the images on both sides of the interface and the contrast between the boundary line and the images on both sides; Step 4, boundary line measurement: the image area processed by HSV superposition and RGB superposition is divided longitudinally, and the water surface boundary line A is used as the low point to obtain slice groups, and each slice is measured separately to calculate the value group; Step 5, dividing line fitting curve: fit the value group into a dividing line array output with the same width as the output image, which is used for water level measurement.

2. A method for determining the boundary between a ship and a water surface according to claim 1, characterized in that: The deep learning algorithm is applied to pre-determine that the current recognition scenario uses a neural network learning algorithm.

3. A method for determining the boundary between a ship and a water surface according to claim 2, characterized in that: The neural network training parameters include the number of iterations: 60,000; the learning rate: 0.2; Weight reduction coefficient: 0.0003, neural network training uses feedforward algorithm for parameter learning.

Images