TWI744795B - Sea level image detection system - Google Patents

Abstract

A sea level image detection system includes an image capturing device, a discrete cosine transformer and a Kalman filter. The image capturing captures a sea level image. The discrete cosine transformer divides the sea level image into a plurality of image blocks and calculates the label values thereof. The discrete cosine transformer compares each label value of the plurality of image blocks with a threshold value to distinguish the sea and the sky, thereby obtaining the discrete sea level image and a sea level. The Kalman filter modifies the discrete sea level image and the sea level based on the past sea level image and the past sea level. By the foregoing configuration, stable sea level images may be obtained while unstable ones may be excluded.

Description

Sea level image detection system

The invention relates to a sea level image detection system that uses a discrete cosine converter and a Kalman filter to obtain a relatively stable sea level image.

In the past, the sea level image detection would be affected by the buoyancy and sinking of the sea, causing the sea level image to shake, drift or tilt. The existing sea-level image detection system uses Hough transform and Kalman Filter to select sea-level images, and filters the sea level images that are swaying, drifting, or tilted; however, sea-level images After the Hough transformation, multiple lines corresponding to the sea level image may be generated, and it is difficult to determine the exact location of the sea level.

The patent of the US Publication No. US20180217243A1 uses ultrasound to detect the height of the sea level, and then knows the sea level position of the ship, but the sea level image has not yet been obtained, and the sea level status of the ship is still unknown.

In summary, the inventor of the present invention thought about and designed a sea level image detection system, in order to improve the lack of conventional technology, and further enhance the application and utilization in the industry.

In view of the above-mentioned conventional problems, the purpose of the present invention is to provide a sea level image detection system to solve the problems faced by the conventional technology.

Based on the above objective, the present invention provides a sea level imaging system, which includes an image extractor, a discrete cosine converter, and a Kalman filter. The image capture device captures sea level images. Discrete The cosine converter is electrically connected to the image capture device to receive the sea level image. The discrete cosine converter divides the sea level image into a plurality of pixel blocks and calculates the corresponding label value. The discrete cosine converter compares the label of each pixel block Value and threshold value to distinguish each pixel block as ocean or sky, and then obtain discrete sea level image and sea level line. The Kalman filter is electrically connected to the discrete cosine converter to receive the discrete sea level image and sea level line. The Kalman filter stores the previous sea level image and the previous sea level line. The Kalman filter is based on the previous sea level image and The previous sea level line corrected the discrete sea level image and sea level line. Through the setting of the discrete cosine converter and Kalman filter, the sea level image is more stable and the sea level image with large sea level fluctuations is removed.

Preferably, each plurality of pixel blocks has a plurality of pixels.

Preferably, the number of the plurality of pixels is different or the same as the number of the plurality of pixel blocks.

Preferably, the discrete cosine converter calculates the DC coefficient values and the plurality of AC coefficient values of the plurality of pixels in each pixel block, and the discrete cosine converter averages the plurality of AC coefficient values of each pixel block to obtain the average coefficient value, and The maximum coefficient value of the average coefficient value of each pixel block is found, and the discrete cosine converter divides the average coefficient value and the maximum coefficient value of each pixel block to obtain the label value.

Preferably, when the label value of the pixel block is less than the threshold value, the pixel block is the sky.

Preferably, when the label value of the pixel block is greater than the threshold value, the pixel block is the ocean.

Preferably, the Kalman filter estimates the current sea level image and the current sea level line based on the previous sea level image and the previous sea level line, and the Kalman filter compares the current sea level image with the current sea level line and the discrete sea. Plane image and sea level line to correct discrete sea level image and sea level line.

Preferably, the Kalman filter estimates the coordinate values corresponding to the current sea level and the wind speed in the current sea level image based on the coordinate value corresponding to the previous sea level line and the wind speed in the previous sea level image.

Preferably, the corrected discrete sea level image is perspectively converted into a stable sea level image based on the corrected sea level line.

Preferably, each pixel block that belongs to the sky near the ocean is connected to each other to form a sea level line.

As mentioned above, the sea level image system of the present invention makes the sea level image more stable through the setting of the discrete cosine converter and the Kalman filter, and removes the sea level image with large sea level fluctuations.

10: Image grabber

20: Discrete Cosine Converter

30: Kalman filter

BSLP: Last sea level image

BSL: Last sea level line

L: marked value

PB: pixel block

SLP: Sea level image

THRS: Threshold

:k狀態的估計值

: Estimated value of k state

:前次狀態的預測結果

: The prediction result of the previous state

B _k μ _{k -1} : Noise in estimation

Pk : The predicted state after estimation

F _k P _{k -1} : The current state of change is estimated k-1 The estimated change time after the previous state

F _k ^T : conversion matrix

Q _k : Filter noise

Figure 1 is a configuration diagram of the sea level imaging system of the present invention.

Figure 2 is a conversion diagram of multiple pixel blocks of the sea level image.

Figure 3 is a schematic diagram of the sea level image after passing through the discrete cosine converter.

Figure 4 is a diagram of the operation mechanism of the Kalman filter.

Figure 5 is a sea level image diagram of the present invention.

The advantages, features, and technical methods of the present invention will be described in more detail with reference to exemplary embodiments and the accompanying drawings to make it easier to understand, and the present invention can be implemented in different forms, so it should not be understood to be limited to what is here. The stated embodiments, on the contrary, are generally known in the technical field For those of you who know, the provided embodiments will make this disclosure more thorough, comprehensive and complete to convey the scope of the present invention, and the present invention will only be defined by the scope of the appended patent application.

It should be understood that although the terms "first", "second", etc. may be used in the present invention to describe various elements, components, regions, layers and/or parts, these elements, components, regions, layers and/or parts Should not be restricted by these terms. These terms are only used to distinguish one element, component, region, layer and/or section from another element, component, region, layer and/or section. Therefore, the "first element", "first part", "first area", "first layer" and/or "first part" discussed below can be referred to as "second element", "second part" , "Second Area", "Second Layer" and/or "Second Part" without departing from the spirit and teachings of the present invention.

In addition, the terms "including" and/or "including" refer to the existence of the features, regions, wholes, steps, operations, elements, and/or components, but do not exclude one or more other features, regions, wholes, steps, operations , The presence or addition of elements, components, and/or combinations thereof.

Unless otherwise defined, all terms (including technical and scientific terms) used in the present invention have the same meanings as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having definitions consistent with their meanings in the context of related technologies and the present invention, and will not be interpreted as idealized or overly formal Unless explicitly defined as such in this article.

Please refer to Figures 1 to 3, which are the layout diagram of the sea level imaging system of the present invention, the schematic diagram of the sea level image after passing through the discrete cosine converter, and the schematic diagram of the sea level image after passing through the discrete cosine converter. As shown in FIGS. 1 to 3, the sea level imaging system of the present invention includes an image extractor 10, a discrete cosine converter 20, and a Kalman filter 30. The image capture device 10 captures the sea level image SLP. The discrete cosine converter 20 is electrically connected to the image capture device 10 to receive the sea level image SLP. The discrete cosine converter 20 divides the sea level image SLP into a plurality of pixel blocks PB and calculates the corresponding label value L. The discrete cosine converter 20 compares the label value L of each pixel block PB and the threshold value THRS to distinguish each pixel block PB. The pixel block PB is ocean SEA or sky SKY, and then the discrete sea level image DSLP and sea level line SL are obtained. The Kalman filter 30 is electrically connected to the discrete cosine converter 20 to receive the discrete sea level image DSLP and the sea level line SL. The Kalman filter 30 stores the previous sea level image BSLP and the previous sea level line BSL, and the Kalman filter 30 Revise the discrete sea level image DSLP and sea level SL based on the previous sea level image BSLP and the previous sea level line BSL. Through the setting of the discrete cosine converter 20 and the Kalman filter 30, the sea level image SLP is more stable, and sea level images with large sea level fluctuations are removed.

Specifically, as shown in FIG. 2, the discrete cosine converter 20 divides the sea level image SLP into a plurality of pixel blocks PB. Each pixel block PB has a plurality of pixels, and the number of the plurality of pixel blocks PB is different Or the same as the number of a plurality of pixels; for example, the discrete cosine converter 20 divides the sea level image SLP into 64 pixel blocks PB, and each pixel block PB has 64 pixels. Of course, it can be based on the sea level image SLP. The actual size is adjusted for the pixel block PB and the number of pixels, and is not limited to the scope listed in the present invention.

，i,j=0,1,2.....,7 其中，P_mn指的是在輸入的像素區塊PB內相對應(m,n)的像素值，每個DCT區塊共有8x8個像素，因此i,j=0~7代表每個像素區塊PB的位置，ε_i ,ε_j是一個變 數係數，唯有在i=j=0時，ε_i ,ε_j設為

，其餘情況(i,j=1~7)，ε _i ,ε _j 則設為

。 Next, the discrete cosine converter 20 calculates the DC coefficient values and the AC coefficient values of the plurality of pixels in each pixel block PB, and its calculation formula is as follows (set the number of pixels to 64, 1 DC coefficient value and 63 AC coefficients value):

， I,j =0 , 1 , 2..... , 7 Among them, P _mn refers to the pixel value corresponding to (m,n) in the input pixel block PB, and each DCT block has a total of 8x8 Pixels, so i , j =0~7 represent the position of each pixel block PB, ε _i , ε _j is a variable coefficient, only when i = j = 0 , ε _i , ε _{j are} set to

, In other cases ( i,j =1~7), ε _i ,ε _j are set to

.

，

為平均係數值，

為平均係數值)，設定門檻值THRS為0.1。 The discrete cosine converter 20 averages the 63 AC coefficient values of each pixel block PB to obtain the average coefficient value, and finds the maximum coefficient value of the average coefficient value of each pixel block PB, and the discrete cosine converter 20 converts each pixel block Divide the average coefficient value and the maximum coefficient value of PB to obtain the label value L (that is,

,

Is the average coefficient value,

Is the average coefficient value), and set the threshold THRS to 0.1.

In addition, as shown in Figure 3, if the label value L of a single pixel block PB is less than the threshold THRS (the value is 0.1), the single pixel block PB is labeled SKY (the white block in Figure 3) ; If the label value L of a single pixel block PB is greater than the threshold value THRS (the value is 0.1), a single pixel block PB is marked as ocean SEA (the black block in Figure 3), and then the SEA that is close to the ocean belongs to the sky Each pixel block PB of SKY is connected to each other as a sea level line SL, so as to obtain a discrete sea level image DSLP and a sea level line SL.

Please refer to Fig. 4, which is a diagram of the operation mechanism of the Kalman filter 30. As shown in Figure 4, the Kalman filter 30 predicts the next state based on the previous state, and its estimation formula is as follows:

為k狀態的估計值(例如速度或座標)，

為前次狀態(k-1狀態)的預測結果，B _k μ _k-1為估計中的雜訊；P _k 是誤差共變異數(亦即誤差值)，主要估算k狀態的不確定性，P _k-1為前次狀態(k-1狀態)的誤差變異數，P _k 透過P _k-1和狀態轉換矩陣(F _k 和F _k ^T )可將不確定性雜訊加入到P _k 的預估，並且由於預測模型也可能會有誤差值，因此額外在加入Q _k (誤差值)，F _k 和F _k ^T 為狀態轉換矩陣，Q _k 為過濾雜訊。 P _k = F _k P _{k -1} F _k ^T + Q _k where,

Is the estimated value of the k state (such as speed or coordinates),

Is the prediction result of the previous state (k-1 state), B _k μ _{k -1} is the noise in the estimation; P _k is the error covariance (that is, the error value), which mainly estimates the uncertainty of the k state, variation of the number of errors P _{k -1} is the previous state of the (k-1 state), the transformation matrix P _k (F _k and F _k ^T) _-1 P _k through state and the noise can be added to the uncertainty of P _k Prediction, and because the prediction model may also have an error value, Q _k (error value) is additionally added, F _k and F _k ^T are state transition matrices, and Q _k is noise filtering.

)，搭配B _k μ _k-1的運算，能取得k狀態的海平面線(

)。 Please refer to Figure 5, which is a sea level image diagram of the present invention. For example, as shown in Figure 5 and in conjunction with Figure 4, the red dots are 2 points of the sea level line SL (this is the sea level line in the k-1 state, which is equivalent to

), with the calculation of B _k μ _{k -1} , the sea level line of state k can be obtained (

).

為估計狀態的變化轉換矩陣，(

+R)^-1為誤差，

為估計變化時間。 In addition, considering the weight value K _k (Kalman gain value), the aforementioned formula is modified as follows:

To estimate the transformation matrix of the state change, (

+ R ) ^-1 is the error,

To estimate the time of change.

包括前次海平面影像BSLP以及前次海平面線BSL所對應的座標值及風速，

為卡爾曼濾波器30所預測出的當前海平面影像以及當前海平面線所對應的座標值及風速，卡爾曼濾波器30比對當前海平面影像和當前海平面線和離散海平面影像DSLP及海平面線SL，以修正離散海平面影像DSLP及海平面線SL，最後修正後離散海平面影像DSLP以修正後海平面線為基準透視轉換為穩定海平面影像。 Apply the aforementioned principles to sea level images,

Including the coordinate value and wind speed corresponding to the previous sea level image BSLP and the previous sea level line BSL,

For the current sea level image predicted by the Kalman filter 30 and the coordinate value and wind speed corresponding to the current sea level line, the Kalman filter 30 compares the current sea level image with the current sea level line and the discrete sea level image DSLP and The sea level line SL uses the modified discrete sea level image DSLP and the sea level line SL, and finally the modified discrete sea level image DSLP uses the modified sea level line as the basis for perspective conversion into a stable sea level image.

It should be mentioned that, in one embodiment, the image capture device 10 is installed in an unmanned ship, and the discrete cosine converter 20 and the Kalman filter 30 are installed in a remote computer; therefore, the image capture device 10 and the discrete Both the cosine converter 20 and the Kalman filter 30 are equipped with corresponding wireless transceivers to transmit the sea level image SLP wirelessly. In another embodiment, the image capture device 10 is installed in an unmanned ship, the discrete cosine converter 20 and the Kalman filter 30 are installed in a remote computer, and it has a cloud server, a cloud server, and an image capture device. 10 and the remote computer network connection; the image capture device 10 sends the sea level image SLP to the cloud server, and the remote computer is connected to the cloud server to obtain the sea level image SLP. The foregoing configuration is only an example, of course, it can also be other preferred configurations, and is not limited to the scope of the present invention.

As mentioned above, the sea level imaging system of the present invention uses the discrete cosine converter 20 and the Kalman filter 30 to stabilize the sea level image SLP and remove the sea level image with large sea level fluctuations. In conclusion, the sea level imaging system of the present invention has the above-mentioned advantages and can obtain a more ideal sea level image.

The above descriptions are merely illustrative and not restrictive. Any equivalent modifications or alterations that do not depart from the spirit and scope of the present invention should be included in the scope of the appended patent application.

10: Image grabber

20: Discrete Cosine Converter

30: Kalman filter

BSLP: Last sea level image

BSL: Last sea level line

DSLP: Discrete sea level image

L: marked value

PB: pixel block

SKY: sky

SEA: Ocean

SL: sea level

SLP: Sea level image

THRS: Threshold

Claims (10)

A sea level image detection system, comprising: an image capture device for shooting a sea level image; a discrete cosine converter electrically connected to the image capture device to receive the sea level image, the discrete cosine converter The sea level image is divided into a plurality of pixel blocks and a corresponding label value is calculated. The discrete cosine converter compares the label value of each pixel block with a threshold value to distinguish each pixel block as ocean or Sky, and then obtain a discrete sea level image and a sea level line; and a Kalman filter electrically connected to the discrete cosine converter to receive the discrete sea level image and the sea level line, the Kalman filter stores a The previous sea level image and a previous sea level line, the Kalman filter corrects the discrete sea level image and the sea level line according to the previous sea level image and the previous sea level line. The sea level image detection system according to claim 1, wherein each of the plurality of pixel blocks has a plurality of pixels. The sea level image detection system according to claim 2, wherein the number of the plurality of pixels is different or the same as the number of the plurality of pixel blocks. According to the sea level image detection system described in claim 2, the discrete cosine converter calculates the DC coefficient value and the plurality of AC coefficient values of the plurality of pixels of each pixel block, and the discrete cosine converter averages each pixel A plurality of AC coefficient values of a block are used to obtain an average coefficient value, and one of the largest coefficient values of the average coefficient value of each pixel block is found, and the discrete cosine converter sums the average coefficient value of each pixel block Divide the maximum coefficient value to obtain the label value. As the sea level image detection system described in claim 1, when the pixel block is The label value is less than the threshold value, and the pixel block is the sky. According to the sea level image detection system of claim 1, when the label value of the pixel block is greater than the threshold value, the pixel block is the ocean. According to the sea level image detection system of claim 1, the Kalman filter estimates a current sea level image and a current sea level line based on the previous sea level image and the previous sea level line, the Kalman filter The filter compares the current sea level image and the current sea level line with the discrete sea level image and the sea level line to correct the discrete sea level image and the sea level line. For the sea level image detection system described in claim 7, the Kalman filter predicts the current sea level line corresponding to the coordinate value corresponding to the previous sea level line and the wind speed in the previous sea level image The coordinate value of and the wind speed in the current sea level image. The sea level image detection system according to claim 1, wherein the corrected discrete sea level image is converted into a stable sea level image by using the corrected sea level line as a reference perspective. The sea level image detection system according to claim 1, wherein each of the pixel blocks that belong to the sky near the ocean is connected to each other to form the sea level line.

Images