Aquaculture monitoring system and method for infaunal bivalves

Abstract

Aquaculture monitoring system is disclosed to address the lack of non-invasive, automated behavioral monitoring technology in the aquaculture of burrowing shellfish. The system includes: an image module for capturing multiple images of the burrowing shellfish's behavioral states; and a processing module including an identification unit, a calculation unit, and an evaluation unit. The identification unit uses an object detection model to label each individual burrowing shellfish as exhibiting behavioral states such as an extended inlet / outlet tube (S) or an exposed shell (E). The calculation unit calculates the number of individuals in the burrowing state (H) based on the known total number of burrowing shellfish. The evaluation unit assesses the aquaculture situation based on the temporal trends of the proportions of various behavioral states. This achieves the effect of non-invasive, automated behavioral monitoring.

Claims

1. An aquaculture monitoring system, specifically for burrowing shellfish, comprising: An image module is used to capture multiple behavioral state images of a burrowing mollusk at time intervals in an aquaculture environment; and a processing module is used to receive and process the multiple behavioral state images captured by the image module. The processing module includes: an identification unit configured to: identify each of the behavioral state images using an object detection model, and label each individual of the burrowing mollusk in each behavioral state image as one of a plurality of behavioral states, including: water inlet / outlet tube protrusion state (S) and shell exposure state (E); a calculation unit configured to: count the number of individuals of the burrowing mollusk in each of the plurality of behavioral states, and calculate the number of individuals of the burrowing mollusk in the burrowing state (H) by subtracting the total number of individuals of the burrowing mollusk in each of the plurality of behavioral states from the known total number of the burrowing mollusk; and an evaluation unit configured to: The aquaculture status of the burrowing shellfish is assessed based on the temporal trends of the proportions of the individual shellfish in the burrowing state (H) and the plurality of behavioral states.

2. As in request item 1, the aquaculture monitoring system, wherein, The identification unit is configured to label each individual of the burrowing shellfish in each of the behavioral state images as one of the following three behavioral states: inlet / outlet tube extended state (S), shell exposed state (E), or inlet / outlet tube extended and shell exposed state (ES); and the calculation unit is configured to count the number of individuals of the burrowing shellfish in each of the behavioral states (S, E, ES), and calculate the number of individuals of the burrowing shellfish in the buried state (H) by subtracting the total number of individuals of the burrowing shellfish in the above behavioral states (S, E, ES) from the known total number of the burrowing shellfish; and the evaluation unit is configured to evaluate the aquaculture status of the burrowing shellfish based on the time trend of the proportion of individuals of the burrowing shellfish in the above four behavioral states (S, E, ES, H).

3. As in request item 1 or 2, an aquaculture monitoring system, wherein, This burrowing mollusk belongs to the Veneridae family.

4. As in request item 3, the aquaculture monitoring system, wherein, The burrowing shellfish is the Taiwan clam (Meretrix taiwanica).

5. As in request item 1 or 2, an aquaculture monitoring system, wherein, The object detection model is trained based on a convolutional neural network.

6. As in request item 5, the aquaculture monitoring system, wherein, This convolutional neural network uses the YOLO architecture.

7. The aquaculture monitoring system as requested in item 1 or 2, wherein, The imaging module includes a webcam.

8. As requested in item 7, the aquaculture monitoring system, wherein, This webcam has both visible light and infrared photography capabilities.

9. The aquaculture monitoring system as requested in item 1 or 2, wherein, The time interval between the capture of the plurality of behavioral state images by the image module is between 1 and 60 minutes.

10. A method for monitoring aquaculture, specifically for monitoring burrowing shellfish, comprising: Multiple behavioral images of this burrowing mollusk were captured at time intervals in the aquaculture environment; An object detection model is used to identify images of various behavioral states, and each individual burrowing mollusk in each image is labeled as one of several behavioral states, including: inlet / outlet tube protrusion state (S) and shell exposure state (E). The number of burrowing mollusks in each of these multiple behavioral states is counted. The number of burrowing mollusks in the burrowing state (H) is calculated by subtracting the total number of burrowing mollusks in each of these multiple behavioral states from the total known number of burrowing mollusks. The aquaculture status of the burrowing mollusks is assessed based on the time trend of the ratio of individuals in the burrowing state (H) to those in each of the multiple behavioral states.

11. As requested in item 10, the aquaculture monitoring method, wherein, Each individual burrowing shellfish in the behavioral state images is labeled with one of the following three behavioral states: inlet / outlet tube extended (S), shell exposed (E), or both inlet / outlet tube extended and shell exposed (ES). The system counts the number of burrowing shellfish in each behavioral state (S, E, ES) and calculates the number of burrowing shellfish in the buried state (H) by subtracting the total number of burrowing shellfish in the above behavioral states (S, E, ES) from the known total number of burrowing shellfish. The farming status of the burrowing shellfish is assessed based on the time trend of the proportion of each individual burrowing shellfish in the above four behavioral states (S, E, ES, H).

12. As requested in items 10 or 11, the aquaculture monitoring methods, wherein, The time interval for capturing the plurality of behavioral state images is between 1 and 60 minutes.

Images