A device for counting still-water fish fry for proliferation and release

Abstract

The present invention discloses a static water fish fry counting device for proliferation and release, which belongs to the technical field of aquaculture. The fish fry counting device is a whole fish fry counting system composed of a fish fry image acquisition integrated box and an algorithm processing integrated box; the fish fry counting device is composed of an algorithm processing integrated box placed on the top surface of the fish fry image acquisition integrated box, the algorithm processing integrated box is composed of a display screen, a camera, a development board, and a power adapter, a fish container is placed at the bottom of the fish fry image acquisition integrated box, a fish fry bag and a fish school partition partition are placed in the fish container, and a fill light and a fill light container are installed under the fish container. The present invention designs a very practical fish fry counting equipment, which has a wide range of use value for proliferation and release activities that require the release of a large number of fish fry.

Description

Technical Field

[0001] The invention belongs to the technical field of aquaculture, and in particular relates to a static water fish fry counting device used for proliferation and release. Background Art

[0002] With the deepening of industrialization and the increase in resource exploitation, fishery resources are becoming increasingly depleted. Faced with the declining quality of fishery products and the increasing demand for aquatic products, as well as the urgent need to maintain the sustainable development of fisheries, how to take effective measures to protect fishery resources and the ecological environment has become an important issue faced by countries with water resources around the world. As one of the management measures widely adopted by fishery management departments, stocking and releasing is of great significance to maintaining the ecological balance of fishery waters and promoting fishery development. This measure refers to the use of artificial methods to directly release or transplant fertilized eggs, larvae or adults of fishery organisms into natural waters such as oceans, beaches, rivers, lakes, and reservoirs to restore and increase the number of populations and improve and optimize the community structure of waters. In a broad sense, it can also include measures to indirectly increase the amount of water population resources, such as improving the ecological environment of waters, placing certain devices (such as egg holders, artificial reefs, etc.) in specific waters, and protecting the reproduction of wild populations, so as to maintain fishery development or restore fishery resources; stocking and releasing is currently a generally accepted method that can ultimately or most directly solve the decline in resource quality.

[0003] Before carrying out the stocking, it is necessary to analyze the specific number of fry to be released based on the estimated biological reproduction rate, size, growth scale of the bait, and the planned catch and supply volume of the area over the years, so as to carry out quantitative release based on the estimated amount. However, the current counting of fry is mainly based on manual counting, which has many limitations. First, manual counting not only requires a lot of time and energy, but also staff operating errors will lead to increased counting errors and cause economic losses. Secondly, aquatic animals such as fish are more sensitive to changes in the external environment and are prone to stress reactions such as swimming, jumping, coma, and loss of appetite, which cause physical damage to the fish body and affect the accuracy of counting. Finally, manual counting makes it feasible for unscrupulous merchants to shortchange fish fry in sales. For example, in the process of weighing and counting, whether the fish size is uniform and whether the fish distribution is evenly mixed when weighing and counting will directly affect the error rate of the final total number.

[0004] In order to solve the drawbacks of manual counting, some scholars have developed a method for counting aquatic animals using sensor counting. Compared with manual counting methods, these counter devices have reduced the damage to aquatic animals and the investment in manpower and material resources to a certain extent, but their installation process is relatively complicated and requires advance investigation of the aquaculture environment. In addition, the counting accuracy is limited by the size of aquatic animals. Therefore, there is an urgent need to use computer vision-based fish fry counting equipment.

[0005] With the development of the times, great progress has been made in hardware equipment such as image acquisition equipment, computer processors, image technology, and deep learning. Computer vision, as a non-invasive technology, uses cameras and computers to simulate the human eye to identify, track, and measure targets. Therefore, it is widely used in all walks of life. Therefore, the automatic fish counting solution based on computer vision has gradually become an important research direction due to its low cost, high precision, and high efficiency.

[0006] However, most of the fish counting algorithms based on computer vision are currently in the laboratory research and innovation stage, and there is still a lack of relevant research on their specific use in real life. Different breeding environments have different restrictions on the application of fish counting algorithms. Therefore, the development of fish counting hardware based on different breeding environments is an indispensable content for the use of fish counting in real life, and the portability, physical stability and adjustability of the equipment in the development of fish counting hardware are all contents that need to be considered. Therefore, for activities that require fry counting for proliferation and release, there is currently a lack of more suitable fry counting equipment, and manual counting methods such as weighing counting and sampling counting are still the main methods. Therefore, the present invention proposes an integrated device based on a counting method of computer vision, the purpose of which is to be applied to the field of proliferation and release to achieve the number of fry in a static water state. Summary of the invention

[0007] The object of the present invention is to provide a static water fry counting device for proliferation and release, characterized in that the fry counting device for proliferation and release is a whole fry counting system composed of a fry image acquisition integrated box and an algorithm processing integrated box; the fry counting device is composed of an algorithm processing integrated box 1 placed on the top surface of a fry image acquisition integrated box 5, a display screen 2 is provided on the front of the algorithm processing integrated box 1, a power outlet 4 is provided on the side, and the outlet of the power outlet 4 is connected to a power adapter 3; wheels 6 are installed on 4 corners of the bottom surface of the fry image acquisition integrated box 5, a fish container 7 is placed at the lower part of the front of the fry image acquisition integrated box 5, a fish school partitioning partition 8 is placed in the fish container 7, a fill light 9 is installed in the fill light container 11, and the fill light container 11 is under the acrylic transparent bottom plate 15 of the fish container; a pull-out handle 10 is provided in front of the fill light container 11 for adjusting the position of the fill light 9; an Nvidia Jetson NX2 is arranged inside the algorithm processing integrated box 1 TX development board 12, a camera 13 is installed on the bottom surface and inserted into the box 5 from the top of the box 5, 14-adjustable pulley 14 is installed on the inner wall of the fry image acquisition integrated box 5 on both sides of the fill light container 11;

[0008] The size of the fish container is designed to be 42*42*4 (cm). According to the requirement that 100 to 300 fry are packed in each bag and the water content of each bag is 5L during the value-added release, when the fry bags are put into the fish container, the water height is only 2 to 3 cm, which can effectively avoid the aggregation and stratification of fish schools. The fish school partitioning partition inside the fish container can directly divide the fish schools into blocks during the process of putting the fry bags into the fish container, thereby reducing clustering.

[0009] The fry image acquisition integrated box is integrated with a light source device and a fish container; the size of the fry image acquisition integrated box is determined according to the distances required from the camera and the light source device to the fish container.

[0010] The algorithm processing integrated box is designed according to the size of a display screen, an adapter, and an Nvidia Jetson NX2 TX development board; a portion of space is reserved on the front of the box to embed the display screen, and the Nvidia Jetson NX2 TX development board and a power adapter are fixed behind the display screen and inside the box; thereby, the operation processing device, the counting effect display device, and the fry image acquisition integrated box are integrated.

[0011] The algorithm processing integrated box embeds a counting program in the Nvidia Jetson NX2 TX development board, which can realize functions such as static image counting, local dynamic video counting and real-time data counting. The development board can collect fish fry image data by importing local data or calling the camera in the integrated box in real time to obtain fish fry images, and perform image preprocessing such as brightness, contrast and smoothness on the collected data through program running to improve the quality of the collected image. Finally, the fish fry counting model embedded in the development board is used to obtain information from the processed image to directly obtain the statistical number of fish fry.

[0012] The algorithm processes the front display in the integrated box, and assists in the display of results after the image data is processed by the Nvidia Jetson NX2 TX development board. The display will output the acquired image information and its corresponding thermal map, as well as the image and counting results after the image preprocessing by the development board.

[0013] The method for counting fish fry for proliferation and release by the static water fish fry counting device for proliferation and release is characterized by comprising the following steps:

[0014] Step 1: According to the requirements of releasing 100 to 300 small and medium-sized fry per bag, inject 5L of seawater and bag the fry;

[0015] Step 2: Put the bagged fish fry into the fish container 7 to spread the fish school evenly to reduce the influence of the fry layering and blocking on the counting result;

[0016] Step 3: Turn on the fill light 9 and test the shooting effect of the camera 13. If the shooting effect is not ideal, adjust the fill light level and the positions of the fish container 7 and the fill light container 11;

[0017] Step 4: Connect the Nvidia Jetson NX2 TX development board to the display power supply, run the counting program, select the import video function, call the camera to obtain the image of the fry in the fish container in real time, and then perform program calculation to obtain the number of fry in the fish container;

[0018] Step 5: Turn off the power of the fill light, development board and display, and push the counting equipment to an idle area to avoid affecting subsequent work;

[0019] In step 3, the light source gear can be adjusted through the fill light controller, and the position of the adjustable pulley under the fish container 7 and the fill light container 11 can be adjusted using an Allen wrench, so that the two containers can be fixed in different positions.

[0020] The beneficial effect of the present invention is that the local structure of the fry counting device is continuously adjusted and refined according to the relevant regulations on fish breeding and release; a very practical fish fry counting equipment is designed, which has a wide range of use value for fish breeding and release activities that require the release of a large number of fish fry. The characteristics of the present invention are as follows:

[0021] 1. The present invention integrates an Nvidia Jetson NX2 TX development board, a display screen, a fill light, a camera and a fish container, and provides a practical device for a fry counting method based on computer vision, so that the counting method studied can be applied in actual use.

[0022] 2. The equipment size is designed based on the release specifications and release conditions of the fry for proliferation and release. Based on this design size, the impact of fry aggregation and stratification on the counting effect can be alleviated.

[0023] 3. After the fry are bagged, the present invention can put the bagged fry in a fish container, call the computer vision-based still water fry counting program in the development board to count the still water fry in a fixed container, reduce physical damage to the outside of the fish body, and increase the survival rate of the fry. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart for the use of a fry counting device for proliferation and release;

[0025] Figure 2 1 is a schematic diagram of the structure of a fry counting device, in which: 1-algorithm processing integrated box, 2-display screen, 3-power adapter, 4-power outlet, 5-fry image acquisition integrated box, 6-wheels, 7-fish container, 8-fish school partition, 9-fill light, 10-pull handle, 11-fill light container;

[0026] Figure 3 yes Figure 2 A side cross-sectional view of the device; in the figure: 12-Nvidia Jetson NX2 TX development board, 13-camera, 14-adjustable pulley, 15-acrylic transparent bottom plate of the fish container.

[0027] Figure 4 This is a schematic diagram of the fry counting effect based on computer vision; DETAILED DESCRIPTION

[0028] The present invention provides a device for counting still-water fry for proliferation and release, which will be further described below in conjunction with the accompanying drawings and embodiments.

[0029] like Figure 2 , Figure 3As shown, the static water fry counting device for proliferation and release is a whole fry counting system consisting of a fry image acquisition integrated box and an algorithm processing integrated box; the fry counting device is composed of an algorithm processing integrated box 1 placed on the top surface of the fry image acquisition integrated box 5, the front of the algorithm processing integrated box 1 is a display screen 2, and a power outlet 4 is arranged on the side, and the outlet of the power outlet 4 is connected to a power adapter 3; wheels 6 are installed on the four corners of the bottom surface of the fry image acquisition integrated box 5, a fish container 7 is placed at the lower part of the front of the fry image acquisition integrated box 5, a fish school partition 8 is placed in the fish container 7, a fill light 9 is installed in the fill light container 11, and the fill light container 11 is under the acrylic transparent bottom plate 15 of the fish container; a pull-out handle 10 is arranged in front of the fill light container 11; it is used to adjust the position of the fill light 9; an Nvidia Jetson NX2 is fixed inside the algorithm processing integrated box 1 TX development board 12, a camera 13 is installed on the bottom surface and inserted into the box 5 from the top, 14-adjustable pulley 14 is installed on the inner wall of the fry image acquisition integrated box 5 on both sides of the fill light container 11; wherein, the wheel 6 is divided into a universal wheel and a brake wheel, the universal wheels are installed at the two corners below the front of the device to facilitate the movement of the device, and the brake wheels are installed at the two corners below the back of the device to facilitate the fixing of the device during the movement;

[0030] The fish container 7 is designed to have a container size of 42*42*4 (cm). According to the document requirements that 100 to 300 fry are packed in bags during value-added release and the water content of each bag is 5L, when the fry bags are inserted into the fish container 7, the water height is only 2 to 3 cm, which can effectively avoid the aggregation and stratification of fish schools. The fish school partitioning baffles inside the fish container 7 can directly divide the fish schools into blocks during the process of inserting the fry bags to reduce clustering.

[0031] The fry image acquisition integrated box is integrated with a light source device and a fish container; the size of the fry image acquisition integrated box is determined according to the distances required from the camera and the light source device to the fish container.

[0032] The algorithm processing integrated box is designed according to the size of a display screen, an adapter, and an Nvidia Jetson NX2 TX development board; a part of the space on the front of the box is embedded with a 13.3-inch touch display screen for counting program call display and counting result display; and the Nvidia Jetson NX2 TX development board and the power adapter are fixed behind the display and inside the box; thereby, the operation processing device, the counting effect display device, and the fry image acquisition integrated box are integrated.

[0033] The algorithm processing integrated box embeds a counting program in the Nvidia Jetson NX2 TX development board, which can realize static image counting, local dynamic video counting, real-time data counting and other functions. The development board can import local data or call the camera in the integrated box in real time to acquire fish fry image data, and perform image preprocessing such as brightness, contrast and smoothness on the acquired data through program running to improve the quality of the acquired image. Finally, the fish fry counting model embedded in the development board is used to acquire information from the processed image, and the statistical number of fish fry can be directly obtained.

[0034] The algorithm processing is integrated into the display in the box. After the Nvidia Jetson NX2 TX development board processes the data, it can assist in the display of the results. The display will output the image information obtained by the camera and its corresponding thermal map, as well as the image and counting results after the image preprocessing by the development board (such as Figure 4 shown).

[0035] Example

[0036] like Figure 1 As shown, the method for counting fish fry for proliferation and release by the static water fish fry counting device for proliferation and release comprises the following steps:

[0037] Step 1: According to the requirements of releasing 100 to 300 small and medium-sized fry per bag, inject 5L of seawater and bag the fry;

[0038] Step 2: Put the bagged fish fry into the fish container 7, the water height in the fish container is only 2-3 cm, the fish school is spread evenly, and the influence of the stratification of the fish fry on the counting result is reduced; the influence of the stratification of the fish fry on the accuracy of the counting method based on computer vision is reduced;

[0039] Step 3: Turn on the fill light 9 and test the shooting effect of the camera 13. If the shooting effect is not ideal, adjust the fill light level and the positions of the fish container 7 and the fill light container 11;

[0040] Step 4: Connect the Nvidia Jetson NX2 TX development board to the display power supply, run the counting program, select the import video function, call the camera to obtain the image of the fry in the fish container in real time, and then perform program calculation to obtain the number of fry in the fish container;

[0041] Step 5: Turn off the power of the fill light, development board and display, and push the counting equipment to an idle area to avoid affecting subsequent work;

[0042] In step 3, the light source gear can be adjusted through the fill light controller, and the adjustable pulley position under the fish container 7 and the fill light container 11 can be adjusted with an Allen wrench, so that the two containers can be fixed in different positions. The fill light 9 is a three-gear adjustable light source, which is placed in the fill light container under the fish container to achieve light source supplement from bottom to top, which to a certain extent avoids the problem of water surface reflection and unclear fry image caused by fill light from top to bottom or from the side, and helps the camera to obtain high-quality fry images;

[0043] In this embodiment, the fill light container 11 is used to place the fill light and its power adapter and other devices;

[0044] In this embodiment, the pull-out handle 10 is mainly used to pull the fish container and the fill-in light container outward, so as to facilitate the insertion of bagged fry into the fish container and the replacement of the fill-in light in the fill-in light container;

[0045] In this embodiment, the Nvidia Jetson NX2 TX development board 12 has 8G video memory, 128G storage space, and a quad-core AEM A57 Complex processor. The development board is used to embed the counting program, implement the running processing of the counting program, and obtain the fry counting result;

[0046] In this embodiment, the camera 13 is a USB high-definition wide-angle camera, which can shoot the entire range of the fish container with a 130-degree distortion-free wide-angle and a focal length of 3 mm, and upload the captured fry images to the counting program for subsequent counting;

[0047] In this embodiment, the main purpose of the adjustable pulley 14 is to help the fish container and the fill light container to be fixed in the integrated box and the two containers to move outward. In addition, in order to still be able to shoot the entire range of the fish container after replacing the camera, the pulley can be adjusted up and down in the box to help the two containers to be fixed at different positions;

[0048] In this embodiment, the acrylic transparent container bottom plate 15 is mainly used as the lower bottom plate of the fish container, so that the fill light below can provide upward fill light to the fry in the fish container to avoid blocking the light source.

[0049] The present invention successfully applies the counting method based on computer vision to the reproduction and release. The fish fry number counting scheme based on the still water fish fry image avoids the advance survey of the aquaculture environment and the complicated and tedious installation process of the equipment compared with other fish fry counting devices based on the breeding environment. Moreover, it only needs to use a camera to shoot the still water fish fry image to obtain the counting result. Compared with other counting devices based on sensors, it causes less damage to the fish body and increases the survival rate of the fish fry during the fish fry counting process.

Claims

1. A device for counting still-water fry for proliferation and release, characterized in that: The fish fry counting device for proliferation and release is a whole fish fry counting system composed of a fish fry image acquisition integrated box and an algorithm processing integrated box. The fish fry counting device comprises: an algorithm processing integrated box (1) is placed on the top surface of the fish fry image acquisition integrated box (5); the front of the algorithm processing integrated box (1) is a display screen (2); a power outlet (4) is arranged on the side; the outlet of the power outlet (4) is connected to a power adapter (3); wheels (6) are installed on the four corners of the bottom surface of the fish fry image acquisition integrated box (5); a fish container (7) is placed on the lower part of the front of the fish fry image acquisition integrated box (5); a fish school partitioning partition (8) is placed in the fish container (7); a fill light (9) is installed in the fill light container (11); the fill light container (11) is under the acrylic transparent bottom plate (15) of the fish container; a pull-out handle (10) is arranged in front of the fill light container (11); a handle for adjusting the position of the fill light (9); an Nvidia Jetson NX2 is arranged inside the algorithm processing integrated box (1). A TX development board (12) is provided with a camera (13) on the bottom surface thereof and is inserted into the image acquisition integrated box (5), and an adjustable pulley (14) is provided on the inner wall of the fry image acquisition integrated box (5) on both sides of the fill light container (11); The size of the fish container is designed to be 42cm*42cm*4cm. According to the document requirement that the fry bags are 100-300 per bag and the water content of each bag is 5L in the value-added release, when the fry bags are put into the fish container, the water height is only 2-3cm, which effectively avoids the aggregation and stratification of the fish school. The fish school partitioning baffle inside the fish container can directly divide the fish school into blocks during the process of putting the fry bags into the fish container, thereby reducing the clustering. The integrated box for obtaining the image of fry is integrated with a light source device and a fish container; the size of the integrated box for obtaining the image of fry is determined according to the distance required from the camera and the light source device to the fish container; The algorithm processing integrated box is designed according to the size of a display screen, an adapter, and an Nvidia Jetson NX2 TX development board; a portion of space is reserved on the front of the box to embed the display screen, and the Nvidia Jetson NX2 TX development board and a power adapter are fixed behind the display screen and inside the box; thereby, the operation processing device, the counting effect display device, and the fry image acquisition integrated box are integrated.

2. The device for counting still-water fry for proliferation and release according to claim 1, characterized in that: The algorithm processing is integrated into a Nvidia Jetson NX2 TX development board in a box and embeds a counting program to realize static image counting, local dynamic video counting and real-time data counting functions. The Nvidia Jetson NX2 TX development board imports local data or calls the camera in the box in real time to acquire fish fry image data, and performs image preprocessing on the acquired brightness, contrast and smoothness data through program running to improve the quality of the acquired image. Finally, the processed image is obtained through a fish fry counting model embedded in the Nvidia Jetson NX2 TX development board to directly obtain the statistical number of fish fry.

3. The device for counting still-water fry for proliferation and release according to claim 1, characterized in that: After the algorithm is integrated into the Nvidia Jetson NX2 TX development board in the box for data processing, the display screen in front of the box will display the results. The display screen will output the acquired image information and its corresponding thermal map, as well as the image and counting results after image preprocessing by the development board.

4. A counting method of the static water fish fry counting device for proliferation and release according to claim 1, characterized in that: The steps include: Step 1: According to the requirements of releasing 100 to 300 small and medium-sized fry per bag, inject 5L of seawater and bag the fry; Step 2: Put the bagged fish fry into the fish container (7) to spread the fish school evenly, so as to reduce the influence of the stratification of the fish fry on the counting result; Step 3: Turn on the fill light (9) and test the shooting effect of the camera (13). If the shooting effect is not ideal, adjust the lighting level of the fill light (9) and the positions of the fish container (7) and the fill light container (11); Step 4: Connect the Nvidia Jetson NX2 TX development board (12) to the display power supply, run the counting program, select the import video function, call the camera (13) to obtain the image of the fry in the fish container (7) in real time, and then calculate the number of fry in the fish container (7) using the program in the development board (12); Step 5: Turn off the fill light (9), Nvidia Jetson NX2 TX development board (12) and the display power, and move the counting device to an idle area to avoid affecting subsequent work.

5. The counting method of the static water fish fry counting device for proliferation and release according to claim 4 is characterized in that: In step 3, the light source gear is adjusted by the fill light controller, and the position of the adjustable pulleys below the fish container (7) and the fill light container (11) is adjusted by using an Allen wrench, so that the two containers can be fixed at different positions.

Images