Intelligent fish returning identification and counting system and method for fishing farm

Abstract

The invention discloses an intelligent identification and counting fish returning system and method for a fishing farm. The system comprises a fish returning box body, a control part, a fixed bracket, a guide part and a hanging protection part, the system adopts a multi-stage guide structure and a comb-shaped bending part, on one hand, the problem of counting accuracy of a traditional single-stage counting structure is solved, on the other hand, small fishes can be distinguished, counting errors are avoided to a great extent, and counting accuracy is remarkably improved. In addition, the invention further provides a counting method based on multistage confirmation time sequence analysis and a recognition algorithm based on feature extraction, whether a counting object is a living object or not can be quickly and accurately recognized through multiple times of confirmation time sequence analysis and the contour-based shape feature and geometric feature extraction technology, and accurate counting is achieved. On the basis, the system can quickly calculate the total weight of the fish according to the technical result, and quickly and safely return the fish to the fish pond, so that secondary damage to the fish body is effectively avoided. Compared with a traditional fish returning method, the system and the method have the advantages that the fish returning accuracy and efficiency of the fishing field can be remarkably improved, the workload of workers in the fishing field is greatly reduced, effective technical support is provided for intelligent management of the fishing field, and the system and the method have remarkable advantages.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent identification and counting returning fish, and in particular to an intelligent identification and counting returning fish system and method for fishing grounds. Background Art

[0002] In today's fast-paced social environment, people's pace of life is accelerating, the time available for leisure and relaxation is becoming increasingly limited, and the ways to relax are relatively scarce. Fishing, as an effective leisure activity to relieve the pressure of fast-paced life, has received widespread attention. With the significant improvement in the living standards of residents, the purpose of fishing has changed from simply obtaining edible fish to pursuing leisure, entertainment and competitive experience. In modern fishing grounds and fishing competitions, it has become a common practice to return the captured fish to the fishing grounds. This measure realizes the recycling of fish resources in the fishing grounds and is in line with the concept of sustainable development.

[0003] In traditional fishing, anglers usually bring their own fish guards or rent traditional fish guards provided by fishing grounds to temporarily store the caught fish. However, as the number of fishing increases, the space inside the fish guards becomes crowded, and fish are easily squeezed and scratched on the surface. In the process of weighing the fish, the traditional method relies on a lot of manpower and material resources, which is not only costly, but also prone to errors, affecting the accuracy of the data. In addition, fishermen bringing their own fish guards will also cause odor residue in the car, poor hygiene, and high cleaning and maintenance costs in the later stage.

[0004] In traditional fishing grounds and competitions, the counting and weighing of fish are mainly done manually, which is inefficient and takes a lot of time and manpower. During manual operation, fish are dehydrated for a long time, and repeated handling and weighing lead to increased fish mortality, further increasing operating costs. In real-life scenarios, for most fishing grounds and competitions where fish of similar size are released, estimating the weight of fish by accurate counting is a simple and effective method. Therefore, achieving accurate counting has become a key link in the fish return process. On the one hand, in order to ensure the accuracy of the counting, a multi-level verification mechanism needs to be adopted; on the other hand, since small fish are inevitably present in the mixed ponds of fishing grounds, electronic fish guards need to have accurate fish size recognition functions to eliminate the interference of small fish on the counting results. In addition, in actual application scenarios, anglers may accidentally mix debris into the fish return system. How to effectively identify and distinguish live fish from debris is also a technical problem that needs to be solved urgently.

[0005] Based on this, the present invention designs an intelligent identification and fish counting system and method for fishing grounds, aiming to solve the pain points in the traditional fishing process and improve the efficiency and experience of fishing activities. Summary of the invention

[0006] The purpose of the present invention is to provide a system and method for intelligently identifying and counting fish returns at fishing grounds, so as to solve the problems raised in the above-mentioned background technology, improve the accuracy and efficiency of fish returns at fishing grounds, automatically recycle fish into fishing ponds, and reduce the workload of staff. In addition, with the help of real-time data transmission and display technology, the number of fish caught by anglers can be displayed in real time, providing anglers and fishing ground managers with instant and accurate data information, and comprehensively improving the experience and management level of fishing activities.

[0007] To achieve the above object, the present invention provides the following technical solutions: An intelligent fish return identification and counting system for fishing grounds, characterized in that it comprises a fish return box, a control component, a fixing bracket, a guide component, and a hanging and protecting component.

[0008] The fish return box is composed of an upper fish protection steel ring, a lower fish protection steel ring, and a waterproof mouth cloth, forming a fish return cavity. The waterproof mouth cloth has a zipper structure, connecting the upper fish protection steel ring and the lower fish protection steel ring. The upper fish protection steel ring and the lower fish protection steel ring are fixed by a left support rod, a right support rod, a front support rod and a fixed bracket. The fish return box is provided with a fish inlet on the top and a fish outlet on the bottom, the fish inlet is connected to the upper fish protection steel ring, and the fish outlet is connected to the lower fish protection steel ring. The fish return box is provided with a guide structure and a counting sensor component inside.

[0009] The control component comprises: MCU control unit; the MCU control unit receives the counting sensor component data and executes the counting algorithm to count the counting objects; Display unit; the display unit is used to display the number of objects to be weighed; Voice unit; the voice unit is used to voice broadcast the quantity information of the weighing object; Communication unit; the communication unit is used to send the number of weighed objects to the server platform for storage; Night light unit: The night light unit is used to provide lighting for anglers when fishing at night; Power supply; the power supply is used to supply power to each unit of the control component; Image sensing component; the image sensing component is used for identifying living fish.

[0010] A control box is installed on the fixed bracket; a power supply and control components are arranged in the control box; different interfaces are arranged on the front, back, left and right sides of the control box for connecting external components with lead-out wires, and a display screen is arranged above the control box for displaying fish protection information.

[0011] The guide component is arranged inside the fish return box, and includes two or more levels of bending components, each level of the guide component includes a guide baffle and a bending component, and the guide baffle is one more than the bending component. The last guide baffle is connected to the fish guard steel ring at the lower end, and the counted objects slide out of the fish return box from the fish guard outlet at the lower end.

[0012] The hanging guard component is fixed on the fish guard steel ring at the lower end and the fixed bracket, and is fixed in the middle by a support rod to make it more stable. The function of the hanging guard component is that when the fishing ground does not want to return the fish caught to the fish pit, the hanging guard can be used to recover the caught fish, and the fish can be pulled away by a fish pulling cart, thereby reducing the workload of fishing.

[0013] Preferably, the fish return box is configured as a receiving box of different shapes, such as a square opening, a round opening, an elliptical opening, etc., with the upper end used for entering fish and the lower end used for exiting fish. The fish return box must be kept level during installation to ensure that the guide baffle and the bending part are tightly combined when the fish are not put in for counting, that is, the counting induction switch is in a closed state. The fish return box is installed on a bracket, and the bracket is fixed to the fishing position by a U-shaped angle iron fixing angle iron.

[0014] Preferably, the angle θ between the guide baffle and the horizontal plane is usually between 30 and 60 degrees, the bending part is composed of a V-shaped baffle, the angle β is usually between 120 and 145 degrees, and the bending part is fixed under the guide baffle by a hinge. The bending part of each level is tightly connected to the guide baffle of the next level, and a counting sensing part is provided at the connection point. The counting sensing part includes a plurality of counting sensors, which collect counting information according to the connection state of the counting sensor received by the guide structure. The counting sensor can be a different type of counting sensing switch, such as a photoelectric switch, a proximity switch, a magnetic induction switch, etc. When the bending part of each level is in a connected state with the guide baffle of the next level, the counting sensing switch is in a closed state; similarly, when the bending part of each level is in a separated state with the guide baffle of the next level, the counting sensing switch is in a disconnected state. The last level bending part adopts a comb-shaped structure, which effectively excludes small fish from the counting objects. .

[0015] Preferably, the image sensing component includes: a linear array CCD camera, a data transmission module, and a data processing unit.

[0016] The invention discloses an intelligent method for identifying and counting returning fish in a fishing ground, which specifically comprises a counting method based on multi-level confirmation time series analysis and an identification algorithm based on feature extraction.

[0017] Preferably, the counting method based on multi-level confirmation time series analysis and the recognition algorithm based on feature extraction, according to the method of multi-level counter sensor time series analysis and judgment and the recognition algorithm of contour feature extraction, realize fast and accurate fish identification and counting.

[0018] Preferably, a counting method based on multi-level confirmation time series analysis and a recognition algorithm based on feature extraction are applied to the above system, and the specific steps are as follows: S401: The intelligent fish return system is powered on, and the fish season information of the current electronic fish guard bound to the fishing ground is obtained from the server, the obtained fish season information is stored in the intelligent fish return system, and the fish guard rental QR code is displayed on the system display screen.

[0019] S402: After the angler scans the QR code on the fish ticket rental interface to bind the fish ticket information and pays the rental fee, the intelligent fish return system authenticates with the server, activates the session for intelligent counting, broadcasts the successful authentication audio, starts timing the session, and initializes the counter i=0.

[0020] S403: The intelligent fish return system detects the counting sensor switches at each level. When all the counting sensor switches are in a closed state, it indicates that the system is normal and fish can be released for counting.

[0021] S404: The angler puts the fish into the fish box through the fish inlet. Under the gravity of the fish, each level of the guide structure opens in sequence, that is, the counting sensor switch of the first level of the guide structure opens first, and then the second level opens, and so on. The time when each level of the counting sensor switch opens is recorded and recorded as TS i1 TS i2 TS i3 .... When the fish enters the next level of guide structure, the counting sensor switch of the previous level of guide structure begins to close, and so on. The time when the counting sensor switch of each level is closed is recorded and recorded as TE i1 TE i2 TE i3 .... At the same time, when the angler puts the fish into the fish return box, the linear array CCD camera 1 placed under the guide baffle 1 quickly captures the image of the fish entering the box; when the fish slides out of the fish return box from the fish outlet, the linear array CCD camera 2 placed under the guide baffle 2 captures the image of the fish leaving the box. The camera continuously shoots at a high frame rate to ensure that every key state of the fish in the entire entry and exit process can be recorded.

[0022] S405: Compare the opening and closing time of each level of counting induction switch, if it meets TS i1 <TS i2 <TS i3 ...、TE i1 <TE i2 <TEi3 ...、TS i1 <TE i1 TS i2 <TE i2 TS i3 <TE i3 ..., execute S600; otherwise, the counting is unsuccessful, return to S403, and count the fish again.

[0023] S406: Execute the recognition algorithm based on feature extraction. If it is identified as a live fish, the counter i=i+1, and the intelligent fish return system displays the count value on the display screen. At the same time, the voice module broadcasts the count and sends the count information to the server platform through the 4G module for storage. Return to S403 and perform the next count of fish return; if it is identified as a foreign object, the count is unsuccessful and returns to S403 for the next count of fish return.

[0024] Compared with the prior art, the present invention has the following beneficial effects: The present invention proposes an intelligent fish returning system for fishing grounds. The system adopts a multi-stage guiding structure and a comb-shaped bending component. On the one hand, it solves the problem of traditional single-stage counting structure in counting accuracy. On the other hand, it realizes the effective distinction between small fish and other non-target objects, avoids misjudgment in the counting process, and significantly improves the reliability of counting.

[0025] The present invention can quickly and accurately identify whether the counting object is a living thing and achieve precise counting through multiple confirmation time series analysis and contour-based shape features and geometric feature extraction technology, and then quickly calculate the total weight of the fish catch. After the counting is completed, the fish can be quickly returned to the fish pond in an efficient and low-damage manner, which greatly reduces secondary damage to the fish, optimizes the operation process of the fishing ground, improves the experience of anglers, and reduces the labor cost of the fishing ground. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0027] Figure 1 A schematic diagram of the external three-dimensional structure of an intelligent identification and counting system for fishing grounds provided by an embodiment of the present invention; Figure 2 A schematic diagram of the internal three-dimensional structure of a system for intelligently identifying and counting fish returns at a fishing ground provided by an embodiment of the present invention; Figure 3is a system block diagram of the control component of the present invention; Figure 4 This is a flow chart of the multi-level confirmation time series analysis and counting method adopted by the present invention.

[0028] Figure 5 This is a schematic diagram of the flow of the recognition algorithm based on feature extraction adopted by the present invention.

[0029] In the accompanying drawings, the components represented by the reference numerals are listed as follows: Fish return box body, 102-control component, 103-fixed bracket, 104-angle iron, 105-hanging protection component, 106-waterproof cloth, 107-upper opening fish inlet, 108-lower opening fish outlet.

[0030] Guide component, 202-control box, 203-display screen, 204-upper fish guard steel ring, 205-lower fish guard steel ring, 206-left support rod, 207-right support rod, 208-front support rod, 209-guide baffle one, 210-guide baffle two, 211-guide baffle three, 212-bending component one, 213-bending component two, 214 counting induction switch one, 215 counting induction switch two, 216 linear array CCD camera one, 217 linear array CCD camera two.

[0031] 301 - MCU control unit, 302 - display unit, 303 - voice unit, 304 - communication unit, 305 - night light unit, 306 - power supply, 307 - counting sensing module one, 308 - counting sensing module two, 309 - image sensing component. DETAILED DESCRIPTION

[0032] 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.

[0033] See also Figure 1-5 , the present invention provides a technical solution: An intelligent fish return identification and counting system for a fishing ground comprises: a fish return box 101, a control component 102, a fixing bracket 103, a guide component 201, and a hanging and protecting component 105.

[0034] The fish return box 101 is composed of an upper fish protection steel ring 204, a lower fish protection steel ring 205, and a waterproof mouth cloth 106, forming a fish return cavity. The waterproof mouth cloth 106 has a zipper structure, connecting the upper fish protection steel ring 204 and the lower fish protection steel ring 205. The upper fish protection steel ring 204 and the lower fish protection steel ring 205 are fixed by a left support rod 206, a right support rod 207, a front support rod 208 and a fixed bracket 103. The fish return box is provided with a fish inlet 107 on the top and a fish outlet 108 on the bottom, the fish inlet is connected to the upper fish protection steel ring 204, and the fish outlet is connected to the lower fish protection steel ring 205. The fish return box is provided with a guide structure 201 and a counting sensor component.

[0035] The control component 102 includes: MCU control unit 301; the MCU control unit 301 receives the counting sensor component data and executes the counting algorithm to count the counting objects; Display unit 302; the display unit 302 is used to display the number of objects to be weighed; Voice unit 303; the voice unit 303 is used to voice broadcast the quantity information of the weighing object; Communication unit 304; the communication unit 304 is used to send the number of weighed objects to the server platform for storage; Night light unit 305; the night light unit 305 is used to provide lighting for anglers when fishing at night; Power supply 306; the power supply 306 is used to supply power to each unit of the control component; Image sensing component 309; the image sensing component 309 is used for identifying living fish.

[0036] A control box 202 is installed on the fixed bracket 103; a power supply 302 and a control component 102 are arranged in the control box 202; different interfaces are arranged on the front, back, left and right sides of the control box 202 for connecting external components with lead-out wires, and a display screen 203 is arranged above the control box for displaying fish protection information.

[0037] The guide component 201 is arranged inside the fish return box 101, and includes two or more levels of guide components 201. Each level of the guide component 201 includes a guide baffle and a bending component, and the guide baffle is one more than the bending component. The last guide baffle is connected to the lower fish guard steel ring 205, and the counted objects slide out of the fish return box 101 from the lower fish guard outlet.

[0038] The hanging guard component 105 is fixed on the lower fish guard steel ring 205 and the fixed bracket 101, and is fixed in the middle by a support rod to make it more stable. The function of the hanging guard component 105 is that when the fishing ground does not want to return the fish to the fish pit, the hanging guard can be used to recover the fish and pull the fish away with a fish pulling cart, thereby reducing the workload of fishing.

[0039] As an embodiment, the fish return box 101 is configured as a receiving box of different shapes, such as a square opening, a round opening, an oval opening, etc., with the upper end used for entering fish and the lower end used for exiting fish. The fish return box 101 must be kept level during installation to ensure that the guide baffle and the bending part are tightly combined when the fish are not put in for counting, that is, the counting induction switch is in a closed state. The fish return box 101 is installed on a bracket, and the bracket is fixed on the fishing position by a U-shaped angle iron 104.

[0040] As an embodiment, the angle θ between the guide baffle and the horizontal plane is usually between 30 and 60 degrees, and is usually 45 degrees. The bending component is composed of a V-shaped baffle, and the angle β is usually between 120 and 145 degrees. The bending component is fixed below the guide baffle by a hinge. The bending component of each level is tightly connected to the guide baffle of the next level, and a counting sensing component is provided at the connection point to determine the switch of the connection. The counting sensing component includes a plurality of counting sensing switches, which collect counting information according to the connection state of the counting sensing switch received by the guide structure. The counting sensing switch can be a sensor of different types, such as a photoelectric switch, a proximity switch, a magnetic induction switch, etc. When the bending component of each level is in a connected state with the guide baffle of the next level, the switch is in a closed state; similarly, when the bending component of each level is in a separated state with the guide baffle of the next level, the switch is in an open state. The last level of the guide bending component adopts a comb-shaped structure, which can effectively distinguish relatively small fish. Usually, the interval between the combs is 50cm~60cm, which effectively excludes small fish from the counting objects. .by Figure 2 Taking the two-stage guide structure as an example, the object to be counted enters the upper opening fish inlet 107 of the fish return box 101, and enters the bending component 1 212 through the guide baffle 1 209 and the guide baffle 2 210. Under the action of gravity, the bending component 1 212 and the guide baffle 2 210 are separated, the door magnetic switch 1 214 is opened, and the object enters the guide baffle 3 211. Similarly, under the action of gravity, the bending component 213 and the guide baffle 3 211 are separated, the counting induction switch 2 215 is opened, and the fish returns to the fishing pond through the fish outlet.

[0041] As an embodiment, the MCU control unit 301 is electrically connected to a power supply 306, and the MCU control unit 301 is electrically connected to the display unit 302, the voice unit 303, the communication unit 304, the night light unit 305, the counting sensor module 1 307, the counting sensor module 2 308, and the image sensor component 309. The power supply 306 can be a separate DC or AC power supply, or a power supply composed of a rechargeable battery and a charger.

[0042] As an embodiment, the image sensing component 309 includes: Linear array CCD camera; There are two linear array CCD cameras, namely, linear array CCD camera 1 216 and linear array CCD camera 2 217, which are respectively placed under guide baffle 1 209 and guide baffle 2 210 to ensure that the camera angle of view can completely cover the fish's entry and exit path. The angle of the linear array CCD camera must be precisely adjusted during installation to ensure that the complete form of the fish can be clearly imaged in the photosensitive area of ​​the linear array CCD camera during the entry and exit process. At the same time, in order to adapt to different ambient lighting conditions, a fill light with adjustable brightness is equipped to work in conjunction with the linear array CCD camera to provide sufficient lighting when the light is dim, ensuring that the collected image is clear and shadow-free.

[0043] Data transmission module: The data interpretation module transmits the image data collected by the linear array CCD camera to the data processing unit through a high-speed data transmission line (such as USB 3.0 or Gigabit Ethernet). In order to ensure the stability and real-time performance of data transmission, the transmission line is reasonably wired to avoid signal interference, and data caching technology is used to temporarily store image data when the data processing unit is busy to prevent data loss.

[0044] Data processing unit; the data processing unit has powerful computing embedded processing capabilities and runs a recognition algorithm based on feature extraction.

[0045] As an embodiment, the counted objects may be fish or other objects.

[0046] The embodiment of the present invention discloses a method for intelligently identifying and counting returning fish at a fishing ground, a counting method based on multi-level confirmation time series analysis and a recognition algorithm based on feature extraction, which are applied to the above-mentioned intelligent counting returning fish system, and the specific steps are as follows: Figure 4 As shown: S401: The intelligent fish return system is powered on, and the fish season information of the current electronic fish guard bound to the fishing ground is obtained from the server, the obtained fish season information is stored in the intelligent fish return system, and the guard rental QR code is displayed on the system display screen.

[0047] S402: After the angler scans the QR code on the fish ticket rental interface to bind the fish ticket information and pays the rental fee, the intelligent fish return system authenticates with the server, activates the session for intelligent counting, broadcasts the successful authentication audio, starts timing the session, and initializes the counter i=0.

[0048] S403: The intelligent fish return system detects the counting sensor switches at each level. When all the counting sensor switches are in a closed state, it indicates that the system is normal and fish can be released for counting.

[0049] S404: The angler puts the fish into the fish box through the fish inlet. Under the gravity of the fish, each level of the guide structure opens in sequence, that is, the counting induction switch 1 214 of the first level of the guide structure opens first, and then the counting induction switch 2 215 of the second level opens, and so on. The time when each level of the counting induction switch is opened is recorded and recorded as TS i1 TS i2 TS i3 .... When the fish enters the next level of guide structure, the counting sensor switch of the previous level of guide structure begins to close, and the time when each level of counting sensor switch is closed is recorded and recorded as TE i1 TE i2 TE i3 .... At the same time, when the angler puts the fish into the fish return box, the linear array CCD camera 216 placed under the guide baffle 209 quickly captures the image of the fish entering the box; when the fish slides out of the fish return box from the fish outlet, the linear array CCD camera 217 placed under the guide baffle 210 collects the image of the fish leaving the box. The camera continuously shoots at a high frame rate to ensure that every key state of the fish in the entire entry and exit process can be recorded.

[0050] S405: Compare the opening and closing time of each level of counting induction switch, if it meets TS i1 <TS i2 <TS i3 ...、TE i1 <TE i2 <TE i3 ...、TS i1 <TE i1 TS i2 <TE i2 TS i3 <TE i3 ..., execute S600; otherwise, the counting is unsuccessful, return to S403, and count the fish again.

[0051] S406: Execute the recognition algorithm based on feature extraction. For specific steps, see Figure 5If it is identified as a live fish, the counter i=i+1, the intelligent fish return system displays the count value on the display screen, and the voice module broadcasts the voice, and sends the count information to the server platform through the 4G module for storage, and returns to S403 for the next count of fish return; if it is identified as a foreign object, the count is unsuccessful, and returns to S403 for the next count of fish return.

[0052] As an embodiment, in step S406, the specific steps of the recognition algorithm based on feature extraction are as follows: S501: Image preprocessing; converting color images into grayscale images to reduce the amount of data for subsequent processing; using filtering algorithms to remove noise from the image, such as Gaussian filtering, median filtering, etc.; normalizing the grayscale value or other feature values ​​of the image to a certain range, such as ([0, 1]) and ([-1, 1]), to eliminate the differences between different images caused by factors such as lighting and shooting conditions, so that subsequent processing is more stable and accurate.

[0053] In step S501, the image grayscale processing is generally based on the conversion formula of the RGB color model, converting the RGB value of each pixel of the color image into a corresponding grayscale value. For example, by using the weighted average method, the calculation formula is: Gray = 0.299R + 0.587G + 0.114B.

[0054] S502: Edge detection; select an edge detection operator. Common edge detection operators include Sobel operator, Prewitt operator, Canny operator, etc.; apply the selected edge detection operator to the preprocessed image to calculate the gradient amplitude and direction of each pixel in the image.

[0055] In step S502, edge points are determined according to the magnitude of the gradient amplitude, a threshold is set, and pixel points with amplitudes greater than the threshold are considered to be edge points, thereby obtaining an edge image.

[0056] S503: Contour extraction; binarization is performed on the edge image to classify the pixels in the image into two categories: edge points and non-edge points; and contour tracing algorithm is used to extract contours from the binarized image.

[0057] In step S504, the binarization process sets a suitable threshold value according to the gradient amplitude or other features obtained by edge detection, sets the pixel points greater than the threshold value as the foreground (edge), and takes the value of 255 or 1; and sets the pixel points less than the threshold value as the background, and takes the value of 0. The contour tracking algorithm can extract the contour from the binary image by using a method based on a boundary chain code or a regional growing method based on a seed point. The method based on the boundary chain code starts from an edge point in the image, and tracks along the edge point in a certain order (such as clockwise or counterclockwise), and records the coordinates and direction of each edge point until it returns to the starting point to form a closed contour; the regional growing method based on the seed point first selects one or more seed points as the starting point of the contour, and then according to certain growth rules, such as the gray value or feature similarity of adjacent pixels, gradually adds the edge points connected to the seed point to the contour until the stop condition is met.

[0058] S504: feature calculation; calculating the shape features (such as perimeter, area, circularity, rectangularity) and geometric features (such as center of gravity coordinates, main axis direction, eccentricity) of the contour, and quantifying the morphological features and spatial distribution characteristics of the contour.

[0059] In step S504, the shape feature calculation is obtained by different calculation methods. The perimeter is obtained by calculating the sum of the distances between all edge points on the contour; the area can be calculated by counting the pixels contained in the contour or using integral images and other methods; the circularity is measured by the ratio of the square of the perimeter to the area, and the closer the ratio is, the closer the contour is to a circle; the rectangularity is expressed by the ratio of the area of ​​the contour to the area of ​​the minimum circumscribed rectangle, reflecting the similarity between the contour and the rectangle. The geometric feature calculation is extracted using corresponding mathematical methods. The centroid coordinates are obtained by calculating the weighted average of the coordinates of all pixels on the contour; the main axis direction is obtained by performing feature decomposition on the covariance matrix of the contour to obtain the eigenvector, where the eigenvector direction corresponding to the larger eigenvalue is the main axis direction; the eccentricity is calculated based on the ratio of the major and minor axes of the elliptical fitting contour, which reflects the degree to which the shape of the contour deviates from a circle.

[0060] S505: Identification decision. Extract contour features from the collected known living creature images to establish a template library and match the contour features of the living creature to be identified with the template library features using methods such as Euclidean distance, cosine similarity, and Mahalanobis distance, and determine identification based on the matching results and a preset threshold.

[0061] In step S505, a large number of images of known living things are collected in advance, and their contour features are extracted according to the above steps to establish a contour feature template library. Each template contains various contour feature values ​​of the corresponding living thing and information such as the category to which it belongs. The contour features of the living thing to be identified are matched with the features in the template library. Common matching methods include Euclidean distance, cosine similarity, Mahalanobis distance, etc. Specifically, Euclidean distance calculates the straight-line distance between two feature vectors, and the smaller the distance, the higher the similarity; cosine similarity measures the cosine value of the angle between two feature vectors, and the closer the value is to 1, the higher the similarity; Mahalanobis distance takes into account the correlation between features and can more accurately measure the similarity of features in some cases. According to the matching results, the template category with the highest similarity is selected as the category of the living thing to be identified. If the matching degree exceeds a certain threshold, the identification is considered successful and the corresponding living thing category is output; if the matching degree is lower than the threshold, it is identified as debris.

[0062] The method of the invention can quickly identify live fish, accurately count them, solve the existing counting problems, and put the counted and identified fish back into the fishing pond.

[0063] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0064] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent fish identification and counting system for fishing grounds, characterized in that: It includes a fish return box, a control component, a fixing bracket, a guide component, and a hanging and protecting component; The fish return box body is composed of an upper fish guard steel ring, a lower fish guard steel ring, and a waterproof mouth cloth to form a fish return cavity; the waterproof mouth cloth has a zipper structure, connecting the upper fish guard steel ring and the lower fish guard steel ring; the upper fish guard steel ring and the lower fish guard steel ring are fixed by a left support rod, a right support rod, a front support rod and a fixed bracket; the fish return box body is provided with a fish inlet on the top and a fish outlet on the bottom, the fish inlet is connected to the upper fish guard steel ring, and the fish outlet is connected to the lower fish guard steel ring; the fish return box body is provided with a guide structure and a counting sensor component; The control component comprises: MCU control unit; the MCU control unit receives the counting sensor component and executes the recognition and counting algorithm, which is specifically a counting method based on multi-level confirmation time series analysis and a recognition algorithm based on feature extraction, and its specific steps are as follows: S401: The intelligent fish return system is powered on, and the fish season information of the current electronic fish guard bound to the fishing ground is obtained from the server, the obtained fish season information is stored in the intelligent fish return system, and the guard rental QR code is displayed on the display screen of the system; S402: After the angler scans the QR code on the fish ticket rental interface to bind the fish ticket information and pays the rental fee, the intelligent fish return system authenticates with the server, activates the session for intelligent counting, broadcasts the authentication success audio, starts the session timing, and initializes the counter i=0; S403: The intelligent fish return system detects the counting induction switches at each level. When all the counting induction switches are in a closed state, it indicates that the system is normal and fish can be released for counting; S404: The angler puts the fish into the fish box through the fish inlet. Under the gravity of the fish, each level of the guide structure opens in sequence, that is, the counting sensor switch of the first level of the guide structure opens first, and then the second level opens, and so on. The time when each level of the counting sensor switch opens is recorded and recorded as TS i1 TS i2 TS i3 ...; When the fish enters the next level of guide structure, the counting induction switch of the previous level of guide structure begins to close, and so on. The time when the counting induction switch of each level is closed is recorded and recorded as TE i1 TE i2 TE i3 ...; At the same time, when the angler puts the fish into the fish return box, the linear array CCD camera 1 placed under the guide baffle 1 quickly captures the image of the fish entering the box; when the fish slides out of the fish return box from the fish outlet, the linear array CCD camera 2 placed under the guide baffle 2 captures the image of the fish leaving the box; the camera continuously shoots at a high frame rate to ensure that every key state of the fish in the entire entry and exit process can be recorded; S405: Compare the opening and closing time of each level of counting induction switch, if it meets TS i1 <TS i2 <TS i3 ...、TE i1 <TE i2 <TE i3 ...、TS i1 <TE i1 TS i2 <TE i2 TS i3 <TE i3 ..., execute S406; otherwise, the counting is unsuccessful, and return to S403 to count the fish again; S406: Execute the recognition algorithm based on feature extraction. If it is identified as a live fish, the counter i=i+1, the intelligent fish return system displays the count value on the display screen, and the voice module broadcasts the voice, and sends the count information to the server platform through the 4G module for storage, and returns to S403 for the next count of fish return; if it is identified as a foreign body, the count is unsuccessful, and returns to S403 for the next count of fish return; Display unit; the display unit is used to display the number of objects to be weighed; Voice unit; the voice unit is used to voice broadcast the quantity information of the weighing object; Communication unit; the communication unit is used to send the number of weighed objects to the server platform for storage; Night light unit: The night light unit is used to provide lighting for anglers when fishing at night; Power supply; the power supply is used to supply power to each unit of the control component; Image sensing component; the image sensing component is used for identifying living fish; A control box is installed on the fixed bracket; a power supply and control components are arranged in the control box; different interfaces are arranged on the front, back, left, and right sides of the control box for connecting external components with lead wires; a display screen is arranged above the control box for displaying fish protection information; The guide component is arranged inside the fish return box, and includes two or more levels of guide components, each level of the guide component includes a guide baffle and a bending component, and the guide baffle is one more than the bending component; the last guide baffle is connected to the fish guard steel ring at the lower end, and the counted objects slide out of the fish return box from the fish guard outlet at the lower end; The hanging guard component is fixed on the fish guard steel ring at the lower end and the fixing bracket, and is fixed in the middle by a supporting rod to make it more stable.

2. The intelligent fish identification and counting system for fishing grounds according to claim 1 is characterized in that: The fish return box is configured as a receiving box body of different shapes, which may be a square opening, a round opening, an elliptical opening, etc. The upper end is used for entering fish, and the lower end is used for discharging fish.

3. The intelligent fish identification and counting system for fishing grounds according to claim 2 is characterized in that: The fish return box must be kept level during installation to ensure that the guide baffle and the bending component are tightly combined when the fish are not released for counting, that is, the counting sensor switch is in a closed state; the fish return box is installed on a bracket, and the bracket is fixed on the fishing position by a U-shaped angle iron.

4. The intelligent fish identification and counting system for fishing grounds according to claim 1 is characterized in that: The angle θ between the guide baffle and the horizontal plane is usually between 30 and 60 degrees. The bending component is composed of a V-shaped baffle, and the angle β is usually between 120 and 145 degrees. The bending component is fixed under the guide baffle by a hinge; the bending component of each level is tightly connected to the guide baffle of the next level, and a counting sensing component is provided at the connection.

5. The intelligent fish identification and counting system for fishing grounds according to claim 4 is characterized in that: The counting sensor component includes a plurality of counting sensors, and the counting information is collected according to the connection state of the counting sensors received by the guide structure; the counting sensors can be different types of counting sensing switches; when the bending component of each level is in a connection state with the guide baffle of the next level, the counting sensing switch is in a closed state; Likewise, when the bending component of each stage is in a separated state from the guide baffle of the next stage, the counting induction switch is in an off state.

6. The intelligent fish identification and counting system for fishing grounds according to claim 1, characterized in that: The last-stage bending component adopts a comb-shaped structure, which effectively excludes small fish from the counting objects.

7. The feature extraction-based recognition algorithm according to claim 1, characterized in that: The specific steps of the recognition algorithm based on feature extraction are as follows: S501: Image preprocessing: converting color images into grayscale images to reduce the amount of data for subsequent processing, using filtering algorithms to remove noise in the image, and normalizing the grayscale value or other feature values ​​of the image to a certain range, such as ([0,1]) and ([-1, 1]), to eliminate differences between different images caused by factors such as lighting and shooting conditions, so that subsequent processing is more stable and accurate; S502: edge detection; selecting an edge detection operator, applying the selected edge detection operator to the preprocessed image, and calculating the gradient magnitude and direction of each pixel in the image; S503: contour extraction; Binarize the edge image and divide the pixels in the image into two categories: edge points and non-edge points; use the contour tracking algorithm to extract the contour from the binary image; S504: feature calculation: calculating the shape features and geometric features of the contour, and quantifying the morphological features and spatial distribution characteristics of the contour; S505: Recognition decision; extract contour features from the collected images of known living things to establish a template library, match the contour features of the living things to be identified with the features of the template library, and determine the recognition based on the matching results and the preset threshold value; collect contour features from the images of known living things to establish a template library, match the contour features of the living things to be identified with the features of the template library, and determine the recognition based on the matching results and the preset threshold value.

Images