Underwater multi-mode detection system for deep and far sea culture

Abstract

The invention discloses an underwater multi-modal detection system for deep and far sea culture, and belongs to the technical field of mariculture, and the underwater multi-modal detection system specifically comprises the following steps: collecting an infrared image in a net cage, and screening a detection area; calculating the activeness of the detection area to obtain an activeness change curve within one day; when the activeness detected in any day is abnormal, counting the proportion of the radiation area of each net cage layer in the whole radiation area, detecting whether the distribution proportion accords with the fish distribution rule of the current time period, and if yes, prompting to check the whole fish school; the decibel value of each layer is collected based on the directional sonar, when the distribution proportion does not conform to the fish distribution rule of the current time period, the radiation area proportion k1 of the abnormal layer and the normal layer is calculated, the sound pressure proportion k2 of the abnormal layer and the normal layer is calculated according to the decibel value, and when the error between the k1 and the k2 is large, the fish school of the abnormal layer is prompted to be checked; according to the invention, multi-mode efficient detection of cultured fish schools is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of marine aquaculture, and in particular to an underwater multimodal detection system for deep-sea aquaculture. Background Art

[0002] With the development of marine aquaculture, deep-sea aquaculture has become an important way to improve aquaculture production and quality. As a high-tech aquaculture method, deep-sea cage aquaculture is a high-investment, high-risk and high-return aquaculture industry type. It has strong wave resistance, high degree of intensiveness, high yield, strong fluidity of aquaculture water, high-quality adult fish, and significant economic benefits. It has become a new economic growth point for expanding living space and developing the marine fishery industry.

[0003] However, the deep sea environment is complex and changeable, which places higher demands on the monitoring and management of aquaculture equipment. Traditional underwater detection systems mostly rely on a single detection method, which has many limitations in practical applications. In addition, existing underwater detection systems often lack in-depth analysis and understanding of the behavior patterns of aquaculture objects, and cannot effectively predict and identify abnormal situations, and thus cannot adjust aquaculture strategies in a timely manner. For example, for certain types of fish such as large yellow croaker, they have different distribution patterns in different time periods. This behavioral characteristic has important guiding significance for aquaculture management and abnormal state detection.

[0004] Therefore, developing an underwater multimodal detection system that can integrate multiple detection methods, accurately analyze the behavior patterns of aquaculture objects, and make intelligent decisions based on them is of great significance for improving the efficiency and safety of deep-sea aquaculture. Summary of the invention

[0005] The purpose of the present invention is to provide an underwater multimodal detection system for deep sea aquaculture to solve the following technical problems:

[0006] Traditional underwater detection systems mostly rely on a single detection method, which has many limitations in practical applications and often lacks in-depth analysis and understanding of the behavioral patterns of aquaculture objects.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] An underwater multimodal detection system for deep sea aquaculture, comprising:

[0009] The infrared imaging module is used to collect infrared images in the cage through a plurality of infrared sensors, wherein the infrared sensors are arranged on the side of the cage, and the radiation areas with radiation intensity greater than a preset threshold are screened from the infrared images and marked as detection areas;

[0010] An image processing module is used to calculate the activity of the detection area every unit time. The activity is the ratio of the pixels that have changed at the current moment relative to the previous moment to all the pixels in the detection area. The activity is counted in units of days and fitted to obtain an activity change curve within a day.

[0011] The layered detection module is used to divide the radiation area collected per unit time into several cage layers from top to bottom when the difference between the activity level detected on any day and the activity level change curve is greater than the preset threshold, and to count the proportion of the radiation area of ​​each cage layer to the total radiation area, and to detect whether the distribution ratio conforms to the fish distribution law of the current period. If so, the staff is prompted to check the overall fish school;

[0012] The sonar detection module is used to collect the decibel value of each cage layer based on directional sonar. When the distribution ratio does not conform to the fish distribution pattern in the current period, the radiation area ratio k1 of the abnormal layer and the normal layer is calculated, and the sound pressure ratio k2 of the abnormal layer and the normal layer is calculated according to the decibel value. When k1-k2 is greater than a, and a is a positive number, the staff is prompted to check the fish schools in the abnormal layer.

[0013] As a further solution of the present invention: the fish distribution pattern is:

[0014] The cages are divided into an upper water layer, a middle water layer and a lower water layer. During the day and low tide, the big yellow croaker will sink to the lower and / or middle water layer, which is the normal stratification for the current period. At night and high tide, the big yellow croaker will actively float to the upper water layer, which is the normal stratification for the current period.

[0015] As a further solution of the present invention: in the image processing module, the activity calculation process is:

[0016] The infrared image is gray-processed to obtain a gray-scale image, the gray-scale image at the current moment is marked as Pi, and the gray-scale image at the previous moment is marked as Pj, and any coordinate point Pi (x, y) in the gray-scale image Pi is selected to obtain the gray-scale value Gi of the point, and the coordinate point Pj (x, y) at the same position in the gray-scale image Pj is selected to obtain the gray-scale value Gj of the point, and the gray-scale difference absolute value d of Gi-Gj is calculated. If d is greater than L, where L is the illumination change error value, then the point Pj (x, y) is a changed pixel point, and if d is less than or equal to L, then the point Pj (x, y) is a unchanged pixel point, and the proportion of all changed pixels to the total pixels of the image is calculated, which is the activity.

[0017] As a further solution of the present invention: in the image processing module, the fitting process of the activity change curve is:

[0018] The outliers of activity are removed based on the median filter, and the activity trend within a single day is fitted by a parabola. During the fitting process, the outliers are further identified and removed through the box plot and Z-score.

[0019] As a further solution of the present invention: in the stratified detection module, when any time period is detected, the proportion of the radiation area of ​​the normal stratification to the total radiation area is detected. If the proportion is greater than or equal to 90%, it complies with the fish distribution law, otherwise it does not comply.

[0020] As a further solution of the present invention: in the sonar detection module, the process of calculating the sound pressure ratio is:

[0021] The decibel value Lp1 of the normal layer and the decibel value Lp2 of the abnormal layer are collected by directional sonar, and the corresponding sound pressure values ​​are calculated according to the following formula:

[0022] P1=P0×10^(Lp1 / 20), P2=P0×10^(Lp2 / 20);

[0023] Where P0 represents the reference sound pressure, then k2=(P2-P x ) / (P1-P x ), P x Indicates the base sound pressure of the current sea area.

[0024] As a further solution of the present invention: it also includes a feeding detection module, when it is necessary to feed bait, obtain the activity and sound pressure value of each layer of the current cage, and according to the weighted average of the activity and sound pressure value of each layer, put a corresponding proportion of bait into each layer.

[0025] Beneficial effects of the present invention:

[0026] The present invention is based on multimodal integration technology, combining infrared imaging, image processing, stratification detection, sonar detection and other means. First, the overall situation of the cage is retrieved through infrared detection, and abnormal conditions in the cage are quickly identified. Then, the activity of the radiation area and the sound pressure value of different stratified areas are analyzed, and combined with the light-avoiding and sound-sensitive habits of large yellow croaker, the stratification and sound pressure intensity of the fish school at different time periods are judged, so as to evaluate the behavior pattern of the fish school and judge whether there is abnormal behavior in the fish school at this time, thereby significantly improving the detection efficiency of fish farming, reducing the detection cost, and assisting the daily management work of the staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention will be further described below in conjunction with the accompanying drawings.

[0028] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] See also Figure 1 As shown, the present invention is an underwater multimodal detection system for deep sea aquaculture, comprising:

[0031] The infrared imaging module is used to collect infrared images in the cage through a plurality of infrared sensors, wherein the infrared sensors are arranged on the side of the cage, and the radiation areas with radiation intensity greater than a preset threshold are screened from the infrared images and marked as detection areas;

[0032] An image processing module is used to calculate the activity of the detection area every unit time. The activity is the ratio of the pixels that have changed at the current moment relative to the previous moment to all the pixels in the detection area. The activity is counted in units of days and fitted to obtain an activity change curve within a day.

[0033] The layered detection module is used to divide the radiation area collected per unit time into several cage layers from top to bottom when the difference between the activity level detected on any day and the activity level change curve is greater than the preset threshold, and to count the proportion of the radiation area of ​​each cage layer to the total radiation area, and to detect whether the distribution ratio conforms to the fish distribution law of the current period. If so, the staff is prompted to check the overall fish school;

[0034] The sonar detection module is used to collect the decibel value of each cage layer based on directional sonar. When the distribution ratio does not conform to the fish distribution pattern in the current period, the radiation area ratio k1 of the abnormal layer and the normal layer is calculated, and the sound pressure ratio k2 of the abnormal layer and the normal layer is calculated according to the decibel value. When k1-k2 is greater than a, and a is a positive number, the staff is prompted to check the fish schools in the abnormal layer.

[0035] The system combines infrared imaging, image processing, layered detection, and sonar detection technology to provide farmers with a comprehensive solution for monitoring the status of aquaculture farms. By analyzing the activity and sound pressure values ​​of the radiation area, the system can evaluate the behavior patterns of fish schools and make reasonable breeding management suggestions based on this; this multimodal integration method enables the system to evaluate the breeding environment from different angles and using different sensor information, ensuring the comprehensiveness and accuracy of the monitoring data.

[0036] In another preferred embodiment of the present invention, the fish distribution pattern is:

[0037] The cages are divided into an upper water layer, a middle water layer and a lower water layer. During the day and low tide, the big yellow croaker will sink to the lower and / or middle water layer, which is the normal stratification for the current period. At night and high tide, the big yellow croaker will actively float to the upper water layer, which is the normal stratification for the current period.

[0038] During the day, yellow croakers tend to sink to the bottom or mid-water layer, especially in the early morning and at low tide. This habit may be related to their aversion to strong light and their preference for turbid water. At night, especially at dusk and high tide, yellow croakers will actively float to the surface. At night, due to the lack of ultraviolet rays, yellow croakers do not need to worry about the risk of fading body color and can float freely.

[0039] Based on the statistical analysis of fish distribution patterns, by identifying the real-time distribution ratio of the current fish school, it is possible to determine whether there is abnormal behavior in the fish school and deal with the abnormal behavior in a timely manner.

[0040] In another preferred embodiment of the present invention, in the image processing module, the activity calculation process is:

[0041] The infrared image is gray-processed to obtain a gray-scale image, the gray-scale image at the current moment is marked as Pi, and the gray-scale image at the previous moment is marked as Pj, and any coordinate point Pi (x, y) in the gray-scale image Pi is selected to obtain the gray-scale value Gi of the point, and the coordinate point Pj (x, y) at the same position in the gray-scale image Pj is selected to obtain the gray-scale value Gj of the point, and the gray-scale difference absolute value d of Gi-Gj is calculated. If d is greater than L, where L is the illumination change error value, then the point Pj (x, y) is a changed pixel point, and if d is less than or equal to L, then the point Pj (x, y) is a unchanged pixel point, and the proportion of all changed pixels to the total pixels of the image is calculated, which is the activity.

[0042] In another preferred embodiment of the present invention, in the image processing module, the fitting process of the activity change curve is:

[0043] The outliers of activity are removed based on the median filter, and the activity trend within a single day is fitted by a parabola. During the fitting process, the outliers are further identified and removed through the box plot and Z-score.

[0044] In another preferred embodiment of the present invention, in the stratified detection module, when any time period is detected, the proportion of the radiation area of ​​the normal stratification to the total radiation area is detected. If the proportion is greater than or equal to 90%, it complies with the fish distribution law, otherwise it does not comply.

[0045] In another preferred embodiment of the present invention, in the sonar detection module, the process of calculating the sound pressure ratio is:

[0046] The decibel value Lp1 of the normal layer and the decibel value Lp2 of the abnormal layer are collected by directional sonar, and the corresponding sound pressure values ​​are calculated according to the following formula:

[0047] P1=P0×10^(Lp1 / 20), P2=P0×10^(Lp2 / 20);

[0048] Where P0 represents the reference sound pressure, then k2=(P2-P x ) / (P1-P x ), P x Indicates the base sound pressure of the current sea area.

[0049] Yellow croakers are very sensitive to sound, especially during the reproductive season. Both male and female fish can make specific sounds. When the fish are densely packed, these sounds are like the sound of boiling water and pine waves. This characteristic is relatively rare among fish.

[0050] Therefore, the present invention detects the activity of the fish school based on this habit and determines whether the fish school is in a normal state.

[0051] In another preferred embodiment of the present invention, it also includes a feeding detection module. When feeding bait is needed, the activity and sound pressure value of each layer of the current cage are obtained, and the corresponding proportion of bait is placed in each layer according to the weighted average of the activity and sound pressure value of each layer.

[0052] Controlling the delivery of bait according to the stratification of the fish schools helps prevent waste of bait and uneven growth of the fish schools.

[0053] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An underwater multimodal detection system for deep sea aquaculture, characterized in that: include: The infrared imaging module is used to collect infrared images in the cage through a plurality of infrared sensors, wherein the infrared sensors are arranged on the side of the cage, and the radiation areas with radiation intensity greater than a preset threshold are screened from the infrared images and marked as detection areas; An image processing module is used to calculate the activity of the detection area every unit time. The activity is the ratio of the pixels that have changed at the current moment relative to the previous moment to all the pixels in the detection area. The activity is counted in units of days and fitted to obtain an activity change curve within a day. The layered detection module is used to divide the radiation area collected per unit time into several cage layers from top to bottom when the difference between the activity level detected on any day and the activity level change curve is greater than the preset threshold, and to count the proportion of the radiation area of ​​each cage layer to the total radiation area, and to detect whether the distribution ratio conforms to the fish distribution law of the current period. If so, the staff is prompted to check the overall fish school; The sonar detection module is used to collect the decibel value of each cage layer based on directional sonar. When the distribution ratio does not conform to the fish distribution pattern in the current period, the radiation area ratio k1 of the abnormal layer and the normal layer is calculated, and the sound pressure ratio k2 of the abnormal layer and the normal layer is calculated according to the decibel value. When k1-k2 is greater than a, and a is a positive number, the staff is prompted to check the fish schools in the abnormal layer.

2. The underwater multimodal detection system for deep sea aquaculture according to claim 1, characterized in that: The fish distribution pattern is: The cages are divided into an upper water layer, a middle water layer and a lower water layer. During the day and low tide, the big yellow croaker will sink to the lower and / or middle water layer, which is the normal stratification for the current period. At night and high tide, the big yellow croaker will actively float to the upper water layer, which is the normal stratification for the current period.

3. The underwater multimodal detection system for deep sea aquaculture according to claim 1, characterized in that: In the image processing module, the activity calculation process is as follows: The infrared image is gray-processed to obtain a gray-scale image, the gray-scale image at the current moment is marked as Pi, and the gray-scale image at the previous moment is marked as Pj, and any coordinate point Pi (x, y) in the gray-scale image Pi is selected to obtain the gray-scale value Gi of the point, and the coordinate point Pj (x, y) at the same position in the gray-scale image Pj is selected to obtain the gray-scale value Gj of the point, and the gray-scale difference absolute value d of Gi-Gj is calculated. If d is greater than L, where L is the illumination change error value, then the point Pj (x, y) is a changed pixel point, and if d is less than or equal to L, then the point Pj (x, y) is a unchanged pixel point, and the proportion of all changed pixels to the total pixels of the image is calculated, which is the activity.

4. The underwater multimodal detection system for deep sea aquaculture according to claim 1, characterized in that: In the image processing module, the fitting process of the activity change curve is: The outliers of activity are removed based on the median filter, and the activity trend within a single day is fitted by a parabola. During the fitting process, the outliers are further identified and removed through the box plot and Z-score.

5. The underwater multimodal detection system for deep sea aquaculture according to claim 2, characterized in that: In the layered detection module, when any time period is detected, the proportion of the radiation area of ​​the normal layer to the total radiation area is detected. If the proportion is greater than or equal to 90%, it is in line with the fish distribution law, otherwise it is not in line with it.

6. The underwater multimodal detection system for deep sea aquaculture according to claim 1, characterized in that: In the sonar detection module, the process of calculating the sound pressure ratio is: The decibel value Lp1 of the normal layer and the decibel value Lp2 of the abnormal layer are collected by directional sonar, and the corresponding sound pressure values ​​are calculated according to the following formula: P1=P0×10^(Lp1 / 20), P2=P0×10^(Lp2 / 20); Where P0 represents the reference sound pressure, then k2=(P2-P x ) / (P1-P x ), P x Indicates the base sound pressure of the current sea area.

7. The underwater multimodal detection system for deep sea aquaculture according to claim 1, characterized in that: It also includes a feeding detection module, which obtains the current cage feed time when feeding is needed. The activity and sound pressure values ​​of each layer are calculated based on the weighted average of the activity and sound pressure values ​​of each layer. Put bait in corresponding proportions in each layer.

Images