Port cargo identification system for roll-on-roll-off ship

Abstract

The invention relates to the technical field of image recognition, in particular to a port cargo recognition system for a roll-on-roll-off ship, which comprises a cargo data acquisition module used for acquiring pre-loaded cargo data, a recognition data acquisition module used for acquiring loading real-time images in real time, and also used for acquiring a ship entering slope angle of the roll-on-roll-off ship and a ship entering speed of loaded cargos in real time, the pre-loading data processing module is used for extracting morphological characteristics of pre-loaded goods and further used for constructing original identification data of the pre-loaded goods, the real-time image analysis module is used for analyzing a profile curve of the loaded goods, and the goods identification module is used for identifying the goods. The identification module is used for matching the original identification data of the pre-loaded goods with the profile curve of the loaded goods and identifying the loaded goods, and the loading management module is used for managing the loading rate of the loaded goods in the next management period. The goods identification efficiency of the roll-on-roll-off ship is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a port cargo recognition system for a roll-on / roll-off ship. Background Art

[0002] With the growth of global trade volume, sea transportation has become one of the main modes of transportation for international trade. Among the many types of ships, ro-ro ships are widely used in the fields of automobiles, machinery and equipment because they can directly load large cargo such as vehicles. However, in actual operations, due to the wide variety and large quantity of cargo, the traditional manual counting and recording methods are inefficient and prone to errors, which brings great challenges to port management. Therefore, the development of an efficient and accurate cargo identification system is of great significance to improving port work efficiency.

[0003] Chinese patent publication number CN109919040A discloses a method and device for identifying the specification information of goods, wherein the method comprises: when a whole pallet of goods enters or leaves the warehouse, controlling a first camera to shoot the top surface of the whole pallet of goods and controlling a second camera to shoot the side surface of the whole pallet of goods, respectively obtaining the top surface image and the side surface image of the whole pallet of goods, wherein the packaging of each of the goods in the whole pallet of goods is the same; respectively segmenting the top surface image and the side surface image to obtain the packaging top surface image and the packaging side surface image of each of the goods; selecting one from each of the packaging top surface images as the target top surface image, and selecting one from each of the packaging side images as the target side surface image; determining the specification information of the whole pallet of goods according to the target top surface image, the target side surface image and the pre-trained recognition model. Thus, the specification information of the whole pallet of goods can be automatically and quickly identified by only obtaining the top surface image and the side surface image of the whole pallet of goods; thus, the invention does not refer to the environmental impact when performing image recognition on the goods, nor does it analyze the deviation between the collected image and the actual goods, and there is a problem of low efficiency in goods recognition. Summary of the invention

[0004] An object of the present invention is to provide a port cargo identification system for a ro-ro ship to solve at least one of the problems existing in the prior art.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A port cargo identification system for a ro-ro ship, characterized by comprising:

[0007] A cargo data acquisition module, used to acquire pre-loaded cargo data;

[0008] Identification data acquisition module, used to collect real-time loading images, and also used to collect the entry angle of the ro-ro ship and the entry speed of the loaded cargo in real time;

[0009] A preloaded data processing module is used to extract the shape features of the preloaded goods according to the shape image of the preloaded goods, and the preloaded processing module is also used to construct original recognition data of the preloaded goods according to the shape features of the preloaded goods, and store the original recognition data;

[0010] A real-time image analysis module is used to analyze the contour curve of the loaded cargo according to the real-time loading image during the monitoring period, and to adjust the analysis process of the contour curve of the loaded cargo according to the ship entry slope angle during the monitoring period;

[0011] A cargo identification module is used to match the original identification data of the pre-loaded cargo with the contour curve of the loaded cargo, identify the loaded cargo according to the matching result, and send the identification result to the user;

[0012] The loading management module is used to manage the loading rate of the loaded goods in the next management period according to the identification result of the loaded goods in the management period.

[0013] Furthermore, the pre-loaded data processing module includes a morphological feature processing unit, and the morphological feature processing unit performs grayscale processing on the pre-loaded cargo morphological image.

[0014] Furthermore, the pre-loaded data processing module also includes a light analysis unit, which is used to compare the cargo reflective intensity b with the preset cargo reflective intensity B to determine the shooting abnormality, which includes normal shooting and abnormal shooting, and adjust Gray (i) to Gray (i)' when the shooting is abnormal, Gray (i) is the gray value of the i-th pixel in the pre-loaded cargo morphology image.

[0015] Furthermore, the preloaded data processing module further includes an edge morphology analysis unit, which is used to perform edge detection on the preloaded cargo image after grayscale processing to obtain edge detection pixels, and use the connection curve of the edge detection pixels as the cargo contour curve;

[0016] The edge morphology analysis unit is used to use the cargo contour curve as original identification data of the pre-loaded cargo.

[0017] Furthermore, the real-time image analysis module includes a real-time contour analysis unit, which is used to perform grayscale processing on the real-time loading images within the monitoring period, and perform contour curve analysis on each grayscale processed real-time loading image to obtain each real-time edge curve;

[0018] The real-time profile analysis unit calculates the real-time line fluctuation α(j) within the monitoring period, and compares the real-time line fluctuation α(j) with the preset fluctuation value A to determine the real-time fluctuation, which includes normal and abnormal. When the real-time fluctuation is normal, E is used as the contour curve of the loaded goods; when the real-time fluctuation is abnormal, the contour curve of the loaded goods is not set, and the abnormal shaking of the goods is sent to the user;

[0019] Wherein, E is the expected contour curve of each real-time edge curve.

[0020] Furthermore, the real-time image analysis module also includes a slope analysis unit, which is used to compare the entry slope angle θ of the ro-ro ship during the monitoring period with each preset slope angle, and when θ is less than a first preset slope angle θ1, adjust the contour curve E of the loaded cargo to E'; when θ is greater than a second preset slope angle θ2, adjust the contour curve E of the loaded cargo to E".

[0021] Furthermore, the real-time image analysis module also includes a ship entry rate analysis unit, which is used to calculate the frame recognition time t within the monitoring period according to the ship entry speed v of the loaded cargo within the monitoring period, and compare the frame recognition time t with the preset time T to determine the rate abnormality, the rate abnormality includes normal and abnormal, and when the rate is abnormal, the first preset slope angle is optimized to θ1', and the second preset slope angle is optimized to θ2'.

[0022] Furthermore, the cargo identification module includes a similarity calculation unit, and the similarity calculation unit is used to calculate the matching similarity γ(r) according to the original identification data of the pre-loaded cargo and the contour curve of the loaded cargo.

[0023] Furthermore, the cargo identification module further includes a cargo identification unit, and the similarity calculation unit is used to compare the calculation result of the matching similarity with the preset similarity Y, and identify the loaded cargo according to the comparison result: if γ(r)<Y, the cargo identification unit determines that the loaded cargo is not similar to the r-th original identification data, and iterates the identification process of the loaded cargo; if γ(r)≥Y, the cargo identification unit determines that the loaded cargo is similar to the r-th original identification data, and sends a recognition success signal to the user;

[0024] The cargo identification unit sends an identification failure signal to the user when the loaded cargo is not similar to all the original identification data.

[0025] Furthermore, the loading management module is used to count the recognition success rate μ2 of the loaded goods within the management cycle, and compare the recognition success rate μ2 with the preset success rate U to determine the recognition abnormality, which includes normal and abnormal, and when the recognition abnormality is abnormal, the loading rate of the loaded goods in the next management cycle is set to V'.

[0026] Compared with the prior art, the beneficial effects of the present invention are: by performing data analysis on pre-loaded cargo to extract the relevant morphology of the pre-loaded cargo, and based on the continuous images of the loaded cargo collected in real time, the same morphological analysis is performed on the loaded cargo, and the interference of the influence of slope, etc. is eliminated, and then the morphological characteristics of the pre-loaded cargo and the loaded cargo are matched and judged, so as to achieve the effect of accurately and efficiently identifying the loaded cargo of the roll-on / roll-off ship. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 Schematic diagram of the structure of the port cargo identification system for ro-ro ships according to this embodiment.

[0029] Figure 2 This is a schematic diagram of the structure of the pre-loaded data processing module in this embodiment.

[0030] Figure 3 Schematic diagram of the structure of the real-time image analysis module of this embodiment.

[0031] Figure 4 Schematic diagram of the structure of the cargo identification module of this embodiment. DETAILED DESCRIPTION

[0032] In order to more clearly illustrate the present invention, the present invention is further described below in conjunction with preferred embodiments and accompanying drawings. Similar components in the accompanying drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the content specifically described below is illustrative rather than restrictive, and should not be used to limit the scope of protection of the present invention.

[0033] It should be noted that, although the terms first, second, third, etc. may be used to describe in the embodiments of the present application, these descriptions should not be limited to these terms. These terms are only used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.

[0034] Specifically, the port cargo identification system for a roll-on / roll-off ship described in this embodiment is applied to the port cargo identification and management of a roll-on / roll-off ship; the port cargo described in this embodiment is vehicles of different models and sizes; the roll-on / roll-off ship described in this embodiment is a small roll-on / roll-off ship; and the port described in this embodiment is an open port.

[0035] See also Figure 1 As shown, it is a schematic diagram of the structure of the port cargo identification system for ro-ro ships in this embodiment, including:

[0036] The cargo data acquisition module is used to acquire pre-loaded cargo data; the pre-loaded cargo data includes a pre-loaded cargo shape image and a pre-loaded cargo weight; the pre-loaded cargo is specifically a vehicle; and the pre-loaded cargo shape image is specifically a top view image reflecting the shape of the pre-loaded cargo.

[0037] Specifically, this embodiment does not specifically limit the method for obtaining the pre-loaded cargo data, and those skilled in the art can freely set it as long as the requirements for obtaining the pre-loaded cargo data are met. In this embodiment, the pre-loaded cargo data is obtained by user interactive input.

[0038] Please continue reading Figure 1 As shown, the system also includes an identification data acquisition module, which is used to collect real-time loading images and the entry slope angle of the roll-on / roll-off ship and the entry speed of the loaded cargo in real time; the real-time loading images include a bird's-eye view of the loaded cargo.

[0039] Specifically, the method for collecting the real-time loading image described in this embodiment is to collect photos through a camera set above the roll-on / roll-off ship of the entry ramp with the photographing direction facing the entry ramp; the method for collecting the entry slope angle of the roll-on / roll-off ship described in this embodiment is to set a slope angle detection sensor on the entry ramp of the roll-on / roll-off ship for real-time collection. The frequency of real-time collection of the real-time loading image described in this embodiment is determined by the hardware performance of the camera, and this embodiment does not make specific limitations. At the same time, the method for obtaining the entry speed of the loaded cargo described in this embodiment is to obtain it in real time by setting a speed sensor at the entrance.

[0040] Please continue reading Figure 1 As shown, the system also includes a pre-loaded data processing module, which is connected to the cargo data acquisition module. The pre-loaded processing module is used to extract the morphological features of the pre-loaded cargo according to the morphological image of the pre-loaded cargo. The pre-loaded processing module is also used to construct original recognition data of the pre-loaded cargo according to the morphological features of the pre-loaded cargo, and store the original recognition data.

[0041] See also Figure 2 As shown, the pre-loaded data processing module includes a morphological feature processing unit, and the morphological feature processing unit is used to perform grayscale processing on the pre-loaded cargo image, so as to perform grayscale processing on the pre-loaded cargo image in a relevant environment, thereby improving the exposure of the grayscale image of the pre-loaded cargo;

[0042] The morphological feature processing unit is used to perform grayscale processing on the pre-loaded cargo morphological image, and the processing process is as follows: the morphological feature processing unit sets the grayscale value of each pixel in the pre-loaded cargo morphological image to Gray(i)=(a2+a3) / 2×R(i)+(a1+a3) / 2×G(i)+(a2+a1) / 2×B(i); wherein R(i) represents the red grayscale value of the i-th pixel of the pre-loaded cargo image, G(i) represents the green grayscale value of the i-th pixel of the pre-loaded cargo image, B(i) represents the blue grayscale value of the i-th pixel of the pre-loaded cargo image, and a1 represents the red weighted weight. , a2 represents the green weighted weight, a3 represents the blue weighted weight, a1+a2+a3=1, a1=BR / (BR+BG+BB), a2=BG / (BR+BG+BB), a3=BG / (BR+BG+BB), BR is the red gray value of the average gray value of the image background, BG is the green gray value of the average gray value of the image background, BB is the blue gray value of the average gray value of the image background; by performing grayscale processing on the pre-loaded cargo shape images under different backgrounds, the processing process for various backgrounds is given, so that the grayscale processing results of the pre-loaded cargo shape images can better reflect the shape of the pre-loaded cargo.

[0043] Please continue reading Figure 2 As shown, the preloaded data processing module also includes a light analysis unit, which is connected to the morphological feature processing unit. The light analysis unit is used to compare the reflection intensity b of the goods with the preset reflection intensity B of the goods, and analyze the shooting abnormality according to the comparison result, so as to reflect the influence of light on the grayscale processing in the grayscale processing process, and then adjust the grayscale processing process of the preloaded goods image according to the analysis result of the shooting abnormality: if b<B, the light analysis unit determines that the shooting is normal and does not make any adjustment; if b≥B, the light analysis unit determines that the shooting is abnormal, and adjusts Gray(i) to Gray(i)', setting Gray(i)'=Gray(i)×ln[e-(bB) / B], where e is the natural logarithm; by numerically comparing the reflection intensity of the preloaded goods morphological image during shooting, and then adjusting the grayscale processing process, when the reflection intensity is higher than the preset value, the image exposure is reduced, so as to better reflect the morphological characteristics of the image.

[0044] Specifically, this embodiment does not specifically limit the value of the preset cargo reflective intensity B, and those skilled in the art can freely set it as long as the value requirement of the preset cargo reflective intensity B is met. In this embodiment, the value of the preset cargo reflective intensity B is 200l x.

[0045] Please continue reading Figure 2As shown, the pre-loaded data processing module further includes an edge morphology analysis unit, and the edge morphology analysis unit is used to analyze the edge curve of the pre-loaded goods according to the pre-loaded goods image after grayscale processing;

[0046] The edge morphology analysis unit is used to perform edge detection on the pre-loaded cargo image after grayscale processing to obtain edge detection pixels, and use the connection curve of the edge detection pixels as the cargo contour curve;

[0047] The edge morphology analysis unit is used to use the cargo contour curve as the original identification data of the pre-loaded cargo; by constructing the original identification data of the pre-loaded cargo, a basic model of identification data is established to provide data and a logical basis for the subsequent identification of the loaded cargo.

[0048] Specifically, in this embodiment, the process of "performing edge detection on the pre-loaded cargo image after grayscale processing" is specifically implemented through the "Canny algorithm"; it can be understood that the data format of the cargo contour curve described in this embodiment is a binarized grayscale image, and the curve formed by the black pixels therein is the cargo contour curve; it is worth noting that the original recognition data format described in this embodiment is specifically a grayscale image.

[0049] Please continue reading Figure 1 As shown, the system also includes a real-time image analysis module, which is connected to the identification data acquisition module. The real-time image analysis module is used to analyze the contour curve of the loaded cargo according to the real-time image loaded during the monitoring period, and adjust the analysis process of the contour curve of the loaded cargo according to the entry slope angle during the monitoring period.

[0050] Specifically, this embodiment does not impose any specific limitation on the value of the monitoring period, and those skilled in the art can freely set it as long as the value requirement of the monitoring period is met. In this embodiment, the monitoring period is set to 3 seconds.

[0051] See also Figure 3 As shown, the real-time image analysis module includes a real-time contour analysis unit, and the real-time contour analysis unit is used to analyze the contour curve of the loaded goods according to the real-time loading image within the monitoring period;

[0052] The real-time contour analysis unit is used to perform grayscale processing on the real-time loading images within the monitoring period, and to perform contour curve analysis on each grayscale processed real-time loading image to obtain each real-time edge curve; the image is uniformly processed through the same processing process, thereby ensuring the consistency of the data;

[0053] The real-time profile analysis unit calculates the real-time line volatility α(j) within the monitoring period and sets

[0054] Where E(k) represents the expected fluctuation of the kth pixel in each real-time edge curve. K represents the number of image pixels of the real-time edge curve, s(j,k) represents the gray value of the kth pixel in the jth real-time edge curve, and J represents the number of real-time images of the load collected during the monitoring period. Set k=1,2...K, j=1,2...J;

[0055] The real-time profile analysis unit compares the real-time line volatility α(j) with the preset volatility value A, and analyzes the real-time volatility within the monitoring period according to the comparison result, and is also used to analyze the contour curve of the loaded goods according to the analysis result of the real-time volatility: if α(j)<A, the real-time profile analysis unit determines that the real-time volatility within the monitoring period is normal, and uses E as the contour curve of the loaded goods; if α(j)≥A, the real-time profile analysis unit determines that the real-time volatility within the monitoring period is abnormal, does not set the contour curve of the loaded goods, and sends a message of abnormal cargo jitter to the user;

[0056] Among them, E is the expected contour curve of each real-time edge curve; by calculating and judging the volatility of the images of the loaded goods collected during the monitoring period, the recognition interference caused by abnormal jittering images is eliminated, thereby improving the efficiency of subsequent loaded goods recognition.

[0057] Specifically, the expected contour curve E of each real-time edge curve described in this embodiment is a set of grayscale values ​​of E(k) pixels; the process of grayscale processing of the real-time loading image in this embodiment is the same as the grayscale processing process of the pre-loaded cargo image, which is not described in detail in this embodiment; at the same time, the process of contour curve analysis of the real-time loading image after grayscale processing in this embodiment is the same as the process of analyzing the edge curve of the pre-loaded cargo according to the pre-loaded cargo image after grayscale processing, which is not described in detail in this embodiment; at the same time, the data format of each real-time edge curve described in this embodiment is the same as that of the cargo contour curve.

[0058] It is understandable that this embodiment does not specifically limit the value of the preset fluctuation value A, and those skilled in the art can freely set it as long as the value requirement of the preset fluctuation value A is met. In this embodiment, the value of the preset fluctuation value A is 0.15.

[0059] Please continue reading Figure 3 As shown, the real-time image analysis module further includes a slope analysis unit, which is connected to the real-time profile analysis unit, and is used to analyze the slope abnormality according to the entry slope angle of the ro-ro ship during the monitoring period, so as to deal with the situation of the reduction of the loading cargo ratio caused by the slope;

[0060] The slope analysis unit compares the entry slope angle θ of the ro-ro ship during the monitoring period with each preset slope angle, and analyzes the slope abnormality according to the comparison result, and is also used to adjust the contour curve of the loaded cargo according to the analysis result of the slope abnormality: if θ≤θ1, the slope analysis unit adjusts the contour curve E of the loaded cargo to E'; if θ1<θ≤θ2, the slope analysis unit determines that the slope during the monitoring period is normal and does not make any adjustment; if θ>θ2, the slope analysis unit adjusts the contour curve E of the loaded cargo to E".

[0061] Wherein, θ1 is the first preset slope angle, θ2 is the second preset slope angle, E' is the result of magnifying the contour curve E of the loaded cargo by p1 times in the direction of entering the ship, and p1=1+(θ1-θ) / θ1 is set; E" is the result of magnifying the contour curve E of the loaded cargo by p2 times in the direction of entering the ship, and p2=1+(θ-θ2) / θ2 is set; by adjusting the contour of the image that exceeds the preset slope to restore the original contour ratio of the image, the recognition efficiency of the loaded cargo is improved.

[0062] Specifically, this embodiment does not impose specific limitations on the values ​​of the first preset slope angle θ1 and the second preset slope angle θ2. Those skilled in the art can set them freely as long as the value requirements of the first preset slope angle θ1 and the second preset slope angle θ2 are met. In this embodiment, the value of the first preset slope angle θ1 is -15°, and the value of the second preset slope angle θ2 is 15°; it is worth noting that the entry slope angle of the roll-on / roll-off ship described in this embodiment is defined as the angle between the sea level and the roll-on / roll-off ship entry ramp. If the angle is a positive number, it indicates that the direction of the entry ramp is above the sea level. If the angle is a negative number, it indicates that the direction of the entry ramp is deep into the sea level.

[0063] Please continue reading Figure 3 As shown, the real-time image analysis module also includes a ship entry rate analysis unit, which is connected to the slope analysis unit. The ship entry rate analysis unit is used to calculate the frame recognition time t within the monitoring period according to the ship entry speed v of the loaded cargo within the monitoring period, and set t=v / S / f; wherein S is the ship entry recognition length of the loaded cargo, and f is the shooting frequency of the loading real-time image within the monitoring period; by calculating the interval time between the shooting of two loading real-time images, the frame recognition time is indirectly determined, which reflects the coupling and correlation of the data;

[0064] The ship entry rate analysis unit is used to compare the frame recognition time t with the preset time T, and analyze the rate abnormality according to the comparison result: if t<T, the ship entry rate analysis unit determines that the rate of loading cargo during the monitoring period is abnormal, and optimizes the first preset slope angle to θ1', sets θ1'=θ1×[1+(Tt) / T], and optimizes the second preset slope angle to θ2', sets θ2'=θ2×[1+(Tt) / T]; if t≥T, the ship entry rate analysis unit determines that the rate of loading cargo during the monitoring period is normal and does not perform optimization; by comparing the frame recognition time length, it is determined whether the image recognition rate is adapted to the loading rate of the loaded cargo, so as to eliminate the occurrence of the phenomenon that the loading rate affects the recognition of the loaded cargo due to the excessive loading rate.

[0065] Specifically, this embodiment does not impose any specific limitation on the value of the preset time length T, and those skilled in the art can freely set it as long as the value requirement of the preset time length T is met. In this embodiment, the optimal value of the preset time length T is 0.1s.

[0066] Please continue reading Figure 1 As shown, the system also includes a cargo identification module, which is connected to the real-time image analysis module and the pre-loaded data processing module. The cargo identification module is used to match the original identification data of the pre-loaded cargo with the contour curve of the loaded cargo, identify the loaded cargo according to the matching result, and send the identification result to the user.

[0067] See also Figure 4 As shown, the cargo identification module includes a similarity calculation unit, which is used to calculate the matching similarity γ(r) according to the original identification data of the pre-loaded cargo and the contour curve of the loaded cargo, and set

[0068] Among them, μ(r) represents the average image grayscale value of the rth original recognition data, set: μE represents the average gray value of the image of the contour curve of the loaded goods, σ(r,E) represents the local covariance between the rth original recognition data and the contour curve of the loaded goods, and the setting σ(r) represents the variance between the image pixels of the rth original recognition data, and is set σ(E) represents the variance between the image pixels of the contour curve of the loaded goods, grey(m,r) represents the grayscale value of the mth pixel of the rth original recognition data, and grey(c,r) represents the grayscale value of the cth pixel of the rth original recognition data. grey(c,E) represents the grayscale value of the cth pixel of the contour curve of the loaded goods. By calculating the matching similarity between the original recognition data and the structure of the contour curve of the loaded goods, the similarity between the two is quantified, which is convenient for subsequent numerical comparison to determine whether the two belong to the same goods, thereby improving the recognition efficiency of the loaded goods.

[0069] Specifically, in this embodiment, the calculation process of μ(r) is the same as that of μE, and the calculation process of σ(r) is the same as that of σ(E). This embodiment does not repeat the calculation process of μE and σ(E).

[0070] Please continue reading Figure 4 As shown, the cargo identification module further includes a cargo identification unit, which is connected to the similarity calculation unit, and the similarity calculation unit is used to compare the calculation result of the matching similarity with the preset similarity Y, and identify the loaded cargo according to the comparison result: if γ(r)<Y, the cargo identification unit determines that the loaded cargo is not similar to the r-th original identification data, and iterates the identification process of the loaded cargo; if γ(r)≥Y, the cargo identification unit determines that the loaded cargo is similar to the r-th original identification data, and sends a recognition success signal to the user;

[0071] When the loaded cargo is not similar to all the original identification data, the cargo identification unit sends an identification failure signal to the user; through iterative judgment of the stored original identification data, accurate identification of the loaded cargo is achieved.

[0072] Specifically, this embodiment does not impose any specific limitation on the value of the preset similarity Y, and those skilled in the art can freely set it as long as the value requirement of the preset similarity Y is met. In this embodiment, the preset similarity Y is set to 0.85.

[0073] Please continue reading Figure 1 As shown, the system further includes a loading management module, the loading management module is connected to the cargo identification module, and the loading management module is used to manage the loading rate of the cargo in the next management cycle according to the identification result of the cargo loaded in the management cycle;

[0074] The loading management module counts the recognition success rate μ2 of the loaded goods within the management cycle, and compares the recognition success rate μ2 with the preset success rate U to determine the recognition abnormality, and is also used to control the loading rate of the loaded goods in the next management cycle according to the comparison result: if μ2<U, the loading management module determines that the recognition is abnormal, and sets the loading rate of the loaded goods in the next management cycle to V', setting V'=V×exp{μ2-U}; if μ2≥U, the loading management module determines that the recognition is normal, and does not change the loading rate of the loaded goods in the next management cycle; wherein V is the loading rate of the loaded goods in the current management cycle; by managing the loading rate of the next management cycle, the result of improving the clarity of the real-time loading image is achieved, thereby improving the accuracy and efficiency of the loaded goods recognition.

[0075] Specifically, this embodiment does not impose any specific limitation on the value of the preset success rate U. Those skilled in the art can set it freely, and it only needs to meet the value requirements of the preset success rate U. In this embodiment, the value of the preset success rate U is 0.95; it can be understood that this embodiment does not impose any specific limitation on the value of the duration of the management cycle. Those skilled in the art can set it freely, and it only needs to meet the value requirements of the duration of the management cycle. In this embodiment, the value of the duration of the management cycle is 10 minutes.

[0076] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A port cargo identification system for ro-ro ships, characterized in that: include: A cargo data acquisition module, used to acquire pre-loaded cargo data; Identification data acquisition module, used to collect real-time loading images, and also used to collect the entry angle of the ro-ro ship and the entry speed of the loaded cargo in real time; A preloaded data processing module is used to extract the shape features of the preloaded goods according to the shape image of the preloaded goods, and the preloaded processing module is also used to construct original recognition data of the preloaded goods according to the shape features of the preloaded goods, and store the original recognition data; A real-time image analysis module is used to analyze the contour curve of the loaded cargo according to the real-time loading image during the monitoring period, and to adjust the analysis process of the contour curve of the loaded cargo according to the ship entry slope angle during the monitoring period; A cargo identification module is used to match the original identification data of the pre-loaded cargo with the contour curve of the loaded cargo, identify the loaded cargo according to the matching result, and send the identification result to the user; The loading management module is used to manage the loading rate of the loaded goods in the next management period according to the identification result of the loaded goods in the management period.

2. The port cargo identification system for ro-ro ships according to claim 1, characterized in that: The pre-loaded data processing module includes a morphological feature processing unit, and the morphological feature processing unit performs grayscale processing on the pre-loaded cargo morphological image.

3. The port cargo identification system for ro-ro ships according to claim 2, characterized in that: The pre-loaded data processing module also includes a light analysis unit, which is used to compare the cargo reflective intensity b with the preset cargo reflective intensity B to determine the shooting abnormality, which includes normal shooting and abnormal shooting. When the shooting is abnormal, Gray (i) is adjusted to Gray (i)', where Gray (i) is the gray value of the i-th pixel point in the pre-loaded cargo morphology image.

4. The port cargo identification system for ro-ro ships according to claim 3, characterized in that: The pre-loaded data processing module further includes an edge morphology analysis unit, which is used to perform edge detection on the pre-loaded cargo image after grayscale processing to obtain edge detection pixels, and use the connection curve of the edge detection pixels as the cargo contour curve; The edge morphology analysis unit is used to use the cargo contour curve as original identification data of the pre-loaded cargo.

5. The port cargo identification system for ro-ro ships according to claim 4, characterized in that: The real-time image analysis module includes a real-time contour analysis unit, which is used to perform grayscale processing on the real-time loading images within the monitoring period, and perform contour curve analysis on each grayscale processed real-time loading image to obtain each real-time edge curve; The real-time profile analysis unit calculates the real-time line fluctuation α(j) within the monitoring period, and compares the real-time line fluctuation α(j) with the preset fluctuation value A to determine the real-time fluctuation, which includes normal and abnormal. When the real-time fluctuation is normal, E is used as the contour curve of the loaded goods; when the real-time fluctuation is abnormal, the contour curve of the loaded goods is not set, and the abnormal shaking of the goods is sent to the user; Wherein, E is the expected contour curve of each real-time edge curve.

6. The port cargo identification system for ro-ro ships according to claim 5, characterized in that: The real-time image analysis module also includes a slope analysis unit, which is used to compare the entry slope angle θ of the ro-ro ship during the monitoring period with each preset slope angle, and when θ is less than a first preset slope angle θ1, adjust the contour curve E of the loaded cargo to E'; when θ is greater than a second preset slope angle θ2, adjust the contour curve E of the loaded cargo to E".

7. The port cargo identification system for ro-ro ships according to claim 6, characterized in that: The real-time image analysis module also includes a ship entry rate analysis unit, which is used to calculate the frame recognition time t within the monitoring period according to the ship entry speed v of the loaded cargo within the monitoring period, and compare the frame recognition time t with the preset time T to determine the rate abnormality, which includes normal and abnormal. When the rate is abnormal, the first preset slope angle is optimized to θ1' and the second preset slope angle is optimized to θ2'.

8. The port cargo identification system for ro-ro ships according to claim 7, characterized in that: The cargo identification module comprises a similarity calculation unit, which is used to calculate the matching similarity γ(r) according to the original identification data of the pre-loaded cargo and the contour curve of the loaded cargo.

9. The port cargo identification system for ro-ro ships according to claim 8, characterized in that: The cargo identification module further includes a cargo identification unit, wherein the similarity calculation unit is used to compare the calculation result of the matching similarity with the preset similarity Y, and identify the loaded cargo according to the comparison result: if γ(r)<Y, the cargo identification unit determines that the loaded cargo is not similar to the r-th original identification data, and iterates the identification process of the loaded cargo; if γ(r)≥Y, the cargo identification unit determines that the loaded cargo is similar to the r-th original identification data, and sends a recognition success signal to the user; The cargo identification unit sends an identification failure signal to the user when the loaded cargo is not similar to all the original identification data.

10. The port cargo identification system for ro-ro ships according to claim 9, characterized in that: The loading management module is used to count the recognition success rate μ2 of the loaded goods within the management cycle, and compare the recognition success rate μ2 with the preset success rate U to determine the recognition abnormality, which includes normal and abnormal, and when the recognition abnormality is abnormal, the loading rate of the loaded goods in the next management cycle is set to V'.

Images