Method for non-destructively detecting degree of erosion of oyster shells by Neslaina larvae

Abstract

The invention discloses a method for non-destructively detecting the degree of erosion of oyster shells by unlady worms, and belongs to the field of marine biological culture disease detection.The method comprises the steps that image collection and preprocessing are conducted, specifically, a Micro CT device is used for scanning shell samples, and Micro CT image data of the shells are obtained; the collected images are preprocessed; data set construction: selecting a part of images from the preprocessed images, labeling by a professional, labeling a U-shaped pipeline and a normal area caused by female insects in the images, and constructing a training data set and a test data set; physical damage to oyster shells caused by traditional shell opening detection is avoided, the risk that the oysters are infected with other germs due to detection operation is remarkably reduced, normal growth and physiological functions of the oysters can be maintained, and the survival rate and the commercial quality of the oysters are improved. In the large-scale production of oyster culture, the economic loss caused by detection can be effectively reduced, and the culture income is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of marine organism aquaculture disease detection, and particularly relates to a method for non-destructively detecting the degree of erosion of oyster shells by Polydora ciliata. Background Art

[0002] As an important marine shellfish aquaculture variety, oysters have important economic value in the aquaculture industry. However, the erosion by Polydora ciliata is a major problem faced by oyster aquaculture. Polydora ciliata will drill holes in the oyster shells, damaging the shell structure, affecting the growth and survival of oysters, reducing the quality and yield of oysters, and causing serious economic losses to farmers. Traditional methods for detecting the degree of erosion of oyster shells by Polydora ciliata, such as direct observation, dissection, etc., not only cause irreversible damage to oysters, but also are difficult to accurately judge the internal erosion situation, and cannot meet the requirements of real-time and accurate detection in large-scale aquaculture. Therefore, it is urgent to develop a method for non-destructively detecting the degree of erosion of oyster shells by Polydora ciliata.

[0003] Currently, the traditional methods for detecting the degree of erosion of oyster shells by Polydora ciliata have significant defects. The direct observation method can only identify obvious erosion marks on the shell surface and is difficult to detect subtle internal damages; although dissection detection can directly show the internal situation, it causes irreparable damage to oysters, cannot achieve continuous monitoring of the same oyster individual, and has low detection efficiency, making it difficult to meet the urgent needs of rapid and accurate detection in large-scale aquaculture scenarios. Therefore, developing a non-destructive, efficient, and accurate detection method has become an urgent problem to be solved in the oyster aquaculture industry. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the purpose of the present invention is to provide a method for non-destructively detecting the erosion of oyster shells by Polydora ciliata, so as to achieve efficient and accurate detection of the degree of erosion of oyster shells by Polydora ciliata without damaging the oysters, and provide strong support for the healthy development of the oyster aquaculture industry.

[0005] Technical Solution: To solve the above technical problems, according to one aspect of the present invention, more specifically, a method for non-destructively detecting the degree of erosion of oyster shells by Polydora ciliata includes the following steps:

[0006] S1. Image acquisition and preprocessing:

[0007] Use a Micro CT device to scan the shell sample to obtain Micro CT image data of the shell;

[0008] Preprocess the collected images;

[0009] S2. Dataset construction:

[0010] Select a part of the preprocessed images and have them annotated by professionals to mark the U-shaped pipes and normal areas caused by Polydora in the images, and construct a training dataset and a test dataset.

[0011] Perform data augmentation operations on the images in the dataset;

[0012] S3. Deep learning model construction and training:

[0013] Construct a deep learning model suitable for Micro CT image detection, such as an architecture based on convolutional neural network (CNN), including convolutional layers, pooling layers, and fully connected layer components.

[0014] Input the training dataset into the constructed deep learning model for training, adopt a suitable loss function (such as cross-entropy loss function) and optimization algorithm (such as stochastic gradient descent algorithm), and continuously adjust the parameters of the model through the backpropagation algorithm until the loss value of the model on the training dataset converges or reaches the preset number of training times.

[0015] S4. Model evaluation and optimization:

[0016] Use the test dataset to evaluate the trained model, calculate evaluation metrics such as the accuracy, recall rate, and F1 value of the model to measure the performance of the model;

[0017] Optimize and adjust the model according to the evaluation results;

[0018] S5. Actual detection application:

[0019] Input the Micro CT image of the oyster to be detected into the optimized deep learning model, and the model outputs the area ratio of the U-shaped pipes invaded by Polydora in the image.

[0020] Furthermore, in the S1 image acquisition and preprocessing step, the preprocessing includes but is not limited to image denoising, gray correction, and image cropping operations to improve the image quality and highlight the feature information related to the U-shaped pipes caused by Polydora.

[0021] 3. A method for non-destructively detecting the degree of erosion of oyster shells by Polydora according to claim 1, wherein: in the S2 dataset construction step, the data augmentation operations include but are not limited to rotation, flipping, and scaling operations to increase the diversity and sample quantity of the dataset and improve the generalization ability of the model.

[0022] Furthermore, in the S4 model evaluation and optimization step, the optimization and adjustment include but are not limited to adjusting the structure of the model, hyperparameters (such as learning rate, batch size, etc.), increasing the amount of training data, and repeating the training and evaluation process until the model performance reaches a satisfactory effect.

[0023] Furthermore, the erosion area calculation formula for the U-shaped pipe is as follows:

[0024] Area of U-shaped pipe = Number of pixel points in the U-shaped pipe area × Physical area corresponding to each pixel point;

[0025] Proportion of U-shaped pipe area = Area of U-shaped pipe ÷ Area of the entire shell.

[0026] The beneficial effects of the method for non-destructively detecting the degree of Polydora ciliata erosion on oyster shells in the present invention are as follows:

[0027] (1) The present invention protects the integrity of oysters: It avoids the physical damage to the oyster shell caused by traditional shell-opening detection, significantly reduces the risk of oysters being infected with other pathogens due to detection operations, helps maintain the normal growth and physiological functions of oysters, and improves their survival rate and commercial quality. In large-scale oyster farming production, this characteristic is particularly crucial, which can effectively reduce economic losses caused by detection and ensure farming benefits.

[0028] (2) The present invention improves detection accuracy: By comprehensively applying a variety of advanced image acquisition technologies (such as high-resolution X-ray imaging) and complex image processing algorithms (including deep feature extraction by deep learning networks, texture feature analysis, shape feature extraction, etc.), it can comprehensively and meticulously capture various changes inside the oyster caused by Polydora ciliata erosion. From microscopic tissue texture changes to macroscopic erosion area shape and distribution, all can be used as the basis for detection. Compared with traditional single-feature detection methods, it greatly improves the accuracy and reliability of judging the degree of Polydora ciliata erosion; Through machine learning or deep learning evaluation models (such as support vector machines, convolutional neural network models, etc.) constructed with a large amount of oyster sample image data with known erosion degrees, after a strict training and optimization process, it can accurately predict the degree of Polydora ciliata erosion on oysters according to the extracted image features. The parameters of the model have been repeatedly debugged and verified to enable it to adapt to the detection requirements of oysters of different varieties, different growth stages, and different environmental conditions, effectively reducing the occurrence of misjudgment and missed judgment situations.

[0029] (3) The present invention realizes a fast detection process: The entire detection process has a high degree of automation. From image acquisition, preprocessing, feature extraction to the final erosion degree evaluation, each link is closely connected and can be efficiently executed through computer programs. Compared with traditional manual shell-opening inspections or semi-automated detection methods based on simple image processing, it greatly shortens the detection time and can detect a large number of oyster samples in a short time, meeting the requirements of real-time monitoring and rapid decision-making in farming production.

[0030] (4) The present invention is applicable to a variety of shellfish species: Since the detection method of the present invention is mainly based on the general characteristics of the internal structure of shellfish and the analysis and processing of image data, and does not depend on the physiological characteristics or anatomical structure differences of specific shellfish species, it can be widely applied to various common shellfish farming species, such as oysters, scallops, mussels, etc. Whether it is marine shellfish or freshwater shellfish, as long as its internal structure can be imaged by non-destructive testing technology, this method can be used to detect the degree of Polydora erosion, with strong generality and adaptability. Brief Description of the Drawings

[0031] The present invention will be further described in detail below with reference to the drawings and specific implementation methods.

[0032] Figure 1 is a schematic flow chart of the present invention;

[0033] Figure 2 is a schematic diagram for calculating the erosion area of the U-shaped pipe in the present invention. Specific Embodiments

[0034] The present invention will be described in detail below with reference to the drawings and embodiments. It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0035] To make the technical solution of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0036] Refer to Figure 1 - Figure 2 , Embodiment

[0037] A method for detecting the degree of Polydora erosion of shellfish without opening the shell includes the following steps:

[0038] Preprocessing of the X-ray image of Polydora - eroded Crassostrea hongkongensis. The X-ray image of the to-be-identified Crassostrea hongkongensis is collected by the small animal in vivo imaging system microCT. Place the Crassostrea hongkongensis sample on the small animal in vivo imaging system microCT instrument, adjust the imaging parameters of the device, and set them as: Voltage (kV): 50, Current (uA): 100, FOV (mm): 86, San Mode: Standard 2s, Stitching Scan: 3, X-ray filter: Cu 0.06 + A10.5.

[0039] Erosion area segmentation. To obtain the characteristic attributes of the U-shaped pipes caused by Polydora infection, it is necessary to exclude the interference with feature extraction, so area segmentation is required, and then the features of the U-shaped pipes can be extracted. U-net performs well on small sample datasets. Its relatively simple structure and fewer parameters make the training process of U-net relatively fast. Therefore, in this case, a U-Net deep learning model improved based on the attention mechanism is used.

[0040] Feature extraction of Polydora's erosion of Crassostrea hongkongensis. It is extracted according to the different shapes of the U-shaped pipes.

[0041] Calculation of the erosion area of the U-shaped pipes: This step aims to quantify the area of the U-shaped pipe region on the shell of Crassostrea hongkongensis caused by Polydora invasion. After image segmentation is completed, the segmented U-shaped pipe region is presented in the form of a binary image, where red represents the U-shaped pipe region and green represents the non-black shell region. Next, by converting the number of pixel points and the actual physical size, the proportion of the black shell region in the total area is calculated.

[0042] The specific calculation formula is: the area of the U-shaped pipe = the number of pixel points in the U-shaped pipe region × the physical area corresponding to each pixel point; the proportion of the U-shaped pipe area = the area of the U-shaped pipe ÷ the area of the entire shell. For example, in two embodiments, through the above formula, the researchers calculated that the proportions of the U-shaped pipe areas on the shells of two sample oysters were 33.18% and 4.34% respectively, as Figure 2 shown.

[0043] Output the calculation result. Based on the above steps, output the area of the U-shaped pipes caused by Polydora's erosion of Crassostrea hongkongensis.

[0044] The method for detecting the degree of Polydora's erosion of shellfish without opening the shell in the present invention avoids many drawbacks of traditional shell-opening detection methods, can efficiently and accurately detect the degree of Polydora's erosion of shellfish, and provides strong technical support for the healthy development of the shellfish aquaculture industry.

[0045] The above-described embodiments only represent several implementation manners of the present invention. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

Claims

1. A method for non-destructive testing of the extent of oyster shell infestation by oyster worms, characterized in that: The following steps are involved: S1. Image acquisition and preprocessing: Scan the shell samples using a Micro CT device to obtain Micro CT image data of the shells; Preprocess the collected images; S2. Dataset construction: A part of the images are selected from the preprocessed images and annotated by professionals to mark the U-shaped pipes caused by insects and normal areas in the images, and to construct training data sets and test data sets. Perform data augmentation operations on the images in the dataset; S3. Deep learning model construction and training: Build a deep learning model suitable for Micro CT image detection, such as a convolutional neural network-based architecture, including convolutional layers, pooling layers, and fully connected layer components. The training data set is input into the constructed deep learning model for training, and a suitable loss function and optimization algorithm are used. The parameters of the model are continuously adjusted through the back propagation algorithm until the loss value of the model on the training data set converges or reaches the preset number of training times. S4. Model evaluation and optimization: Use the test data set to evaluate the trained model and calculate the model's accuracy, recall, F1 value and other evaluation indicators to measure the model's performance; According to the evaluation results, the model is optimized and adjusted; S5. Actual detection application: The Micro CT image of the oyster to be tested is input into the optimized deep learning model, and the model outputs the area ratio of the U-shaped pipe infected by the nymph in the image.

2. A method for non-destructive detection of the extent of oyster shell infestation by insects according to claim 1, characterized in that: In the S1 image acquisition and preprocessing step, the preprocessing includes but is not limited to image denoising, grayscale correction, and image cropping operations to improve image quality and highlight feature information related to the U-shaped pipe caused by the talented lady bug.

3. A method for non-destructive detection of the extent of oyster shell infestation by oyster insects according to claim 1, characterized in that: In the S2 dataset construction step, the data augmentation operations include but are not limited to rotation, flipping, and scaling operations to increase the diversity and sample quantity of the dataset and improve the generalization ability of the model.

4. A method for non-destructive detection of the extent of oyster shell infestation by oyster insects according to claim 1, characterized in that: In the S4 model evaluation and optimization step, the optimization adjustment includes but is not limited to adjusting the structure and hyperparameters of the model, increasing the amount of training data, and repeating the training and evaluation process until the model performance achieves satisfactory results.

5. The method for non-destructive detection of the extent of oyster shell infestation by oyster insects according to claim 1, characterized in that: The calculation formula for the erosion area of ​​the U-shaped pipe is: U-shaped pipe area = number of pixels in the U-shaped pipe area × physical area corresponding to each pixel; U-shaped pipe area ratio = U-shaped pipe area ÷ total shell area.

Images