Fish Detection Research Articles

Population-level shape variation and otolith asymmetry in Diplodus annularis

Sound detection in fishes relies on the inner ear and peripheral structures, such as calcareous otoliths, which play a crucial role in perceiving movement, orientation, and balance. Otoliths, in particular, respond to various environmental factors including temperature, salinity, and food availability, making them valuable indicators of ecological conditions. This study applies geometric morphometrics (GMM) to analyze the otolith shape of Diplodus annularis (Linnaeus, 1758) from two distinct populations located in the Gulf of Asinara (Porto Torres, Sardinia) and the northern Adriatic Sea (Le Tegnue). By using GMM techniques, precise and quantifiable differences in otolith morphology were revealed between the two populations, demonstrating significant shape variation. In addition, fluctuating asymmetry (FA), which serves as a marker for developmental stability and environmental stress, was assessed. The results show marked disparities in FA between the populations: samples from Porto Torres exhibited posterior asymmetry, while samples from Le Tegnue showed asymmetry in the anterior region of the otoliths. These findings underscore the sensitivity of GMM in detecting even subtle morphological differences, making it a powerful tool for studying environmental and genetic influences on marine species.

DeepFins: Capturing dynamics in underwater videos for fish detection

DeepFins: Capturing dynamics in underwater videos for fish detection

Fish Catch Sorting and Detection Model Improved Based on YOLOv8 Model

The primary challenge in trawl fishing lies in its limited selectivity, resulting in highly diverse fish catches with severe mixed-species issues. The catches from trawl fishing require manual sorting, leading to low work efficiency and high labor demands. In order to tackle this issue, the present study introduces an enhanced version of the DWR module by utilizing DilatedReparamBlock convolution, introducing a novel dilated convolution module. This module is integrated into the YOLOv8 model, enhancing its capacity to capture features from the enlarged receptive field in the upper network layers. Furthermore, a new attention mechanism based on a multi-branch structure with the CA attention mechanism is incorporated into the YOLOv8 model. This attention mechanism fully extracts image features, enhances feature representation, strengthens the generation of offsets and sampling weights, and improves the accuracy of target recognition. Lightweight improvements to the detection head are achieved through the use of shared convolutions, ultimately resulting in a significant reduction in the number of model parameters. Our empirical findings indicate that the refined model exhibits a 2.2% enhancement in mAP@0.5 when benchmarked against the initial YOLOv8 model, Offering a significant reference for the advancement of an effective embedded system for fish sorting.

A lightweight model for echo trace detection in echograms based on improved YOLOv8

With the rise of underwater unmanned platforms like unmanned boats, ROVs, and AUVs, there’s an increasing need for underwater detection technologies. Researchers have merged scientific echosounders with these platforms for biometric applications. However, current detection models are too parameter-heavy to embed in echosounders and struggle with noisy, irregular, and dense echograms. This paper introduces YOLOv8-SBE, a lightweight fish detection model based on YOLOv8, addressing these issues by enhancing feature extraction, information fusion, and small object recognition. YOLOv8-SBE adds the C2f_ScConv module to improve efficiency and reduce parameters, incorporates the BiFPN structure to enhance information transfer, and uses the EMA attention module for better small target recognition. It reduces computational complexity by 18.5%, decreases model parameters by 40%, and improves mAP0.5 to 79.5% and mAP0.5:0.95 to 58.2%, making it suitable for echosounders with limited resources.

Comparison of YOLOv7, YOLOv8, and YOLOv9 for Underwater Coral Reef Fish Detection

Comparison of YOLOv7, YOLOv8, and YOLOv9 for Underwater Coral Reef Fish Detection

Fish-Finder: A robust small target detection method for aquaculture fish in low-quality underwater images.

Underwater fish object detection serves as a pivotal research direction in marine biology, aquaculture management, and computer vision, yet it poses substantial challenges due to the complexity of underwater environments, occultations, and the small-sized and frequently moving fish in aquaculture. Addressing these challenges, we propose a novel underwater fish object detection algorithm named Fish-Finder. First, we engendered a structure titled "C2fBF," utilizing the dual-path routing attention protocol of BiFormer. The primary objective of this structure is to alleviate the perturbations induced by underwater intricacies during the phase of downsampling in the backbone network, thereby discerning and conserving finer contextual features. Subsequently, we co-opted the RepGFPN method within our neck network-a distinctive approach that adeptly merges high-level semantic constructs with low-level spatial specifics, thus fortifying its multi-scale detection prowess. Then, in an endeavor to diminish the sensitivity toward positional aberrations during the detection of diminutive aquatic creatures, we incorporated a novel bounding box regression loss function, the Wasserstein loss, to the existing CIoU. This innovative function gauges the congruity between the predicted bounding box Gaussian distribution and the reference bounding box Gaussian distribution. Finally, in regard to the dataset, we independently assembled a specific dataset termed "SmallFish." This unique dataset, meticulously designed for the detection of small-scale fish within intricate underwater settings, includes 5000 annotated images of small fish. Experimental results demonstrate that, compared to the state-of-the-art detection methods, our proposed method improves the accuracy by and , and mean average precision (mAP) increases and in public dataset Kaggle-Fish and our SmallFish dataset, respectively.

P-type accumulation mode organic electrochemical transistor biosensor for xanthine detection in fish

Food waste is a global challenge that needs to be mitigated in the development of more sustainable societies. From manufacturers to customers, food biosensors could effectively reduce the amount of discarded food and provide more precise predictions of freshness with respect to pre-decided expiration dates. In this study, we developed a novel organic electrochemical transistor (OECT)-based xanthine biosensor. The OECT-based biosensor is based on the p-type conjugated polymer, p(g42T-TT) as the channel, and incorporated xanthine oxidase (XOD) as the biorecognition element. The OECT thus acts as a transducer and amplifier of the enzymatic oxidation of xanthine. Real-time monitoring of xanthine using the OECT-based biosensor led to a linear range between 5 and 98 μM (R2=0.989), 3.28 μM limit of detection, and high sensitivity up to 21.8 mA/mM. Real sample tests showed that the biosensor can detect the accumulation of xanthine in fish meat from 0 to 6 days of degradation. Interference tests with ascorbic acid and uric acid and spike-and-recovery tests with fish samples indicated that as-designed biosensors have good selectivity and accuracy. The developed biosensors show great potential for point-of-care testing applied to food monitoring.

Hyperspectral imaging and deep learning for parasite detection in white fish under industrial conditions

Parasites in fish muscle present a significant problem for the seafood industry in terms of both quality and health and safety, but the low contrast between parasites and fish tissue makes them exceedingly difficult to detect. The traditional method to identify nematodes requires removing fillets from the production line for manual inspection on candling tables. This technique is slow, labor intensive and typically only finds about half the parasites present. The seafood industry has struggled for decades to develop a method that can improve the detection rate while being performed in a rapid, non-invasive manner. In this study, a newly developed solution uses deep neural networks to simultaneously analyze the spatial and spectral information of hyperspectral imaging data. The resulting technology can be directly integrated into existing industrial processing lines to rapidly identify nematodes at detection rates (73%) better than conventional manual inspection (50%).

Improving CNN Fish Detection and Classification with Tracking

The regular and consistent monitoring of marine ecosystems and fish communities is becoming more and more crucial due to increasing human pressures. To this end, underwater camera technology has become a major tool to collect an important amount of marine data. As the size of the data collected outgrew the ability to process it, new means of automatic processing have been explored. Convolutional neural networks (CNNs) have been the most popular method for automatic underwater video analysis for the last few years. However, such algorithms are rather image-based and do not exploit the potential of video data. In this paper, we propose a method of coupling video tracking and CNN image analysis to perform a robust and accurate fish classification on deep sea videos and improve automatic classification accuracy. Our method fused CNNs and tracking methods, allowing us to detect 12% more individuals compared to CNN alone.

Optimizing Formalin Detection in Fish Using QCM Sensors with TOMAC Membrane Coatings for Product Quality Monitoring

<span lang="EN-AU">Detection of formalin in fishery products is a significant concern in the food industry to ensure consumer safety. This study compared the performance of Quartz Crystal Microbalance (QCM) sensors without a membrane and with a Trioctyl methyl ammonium Chloride (TOMAC) membrane coating in detecting formalin in fish samples. The research findings indicate that QCM without a membrane for formalin samples has a lower detection limit of 150 ppm and an upper detection limit of 350 ppm with a sensitivity of 2194.171 Hz/ppm. On the other hand, QCM with a TOMAC membrane coating has a lower detection limit of 400 ppm and an upper detection limit of 550 ppm with a sensitivity of 842.7551 Hz/ppm. Meanwhile, QCM without a membrane for formalin in fish samples has a lower detection limit of 450 ppm and an upper detection limit of 650 ppm with a sensitivity of 15386.38 Hz/ppm. At the same time, QCM with a TOMAC membrane coating for formalin in fish samples has a lower detection limit of 350 ppm and an upper detection limit of 500 ppm with a sensitivity of 23108.9 Hz/ppm. Response time analysis shows that both sensors reach a steady state condition after 12 seconds. This study highlights the importance of selecting appropriate sensors for detecting formalin in fishery products, considering detection limits, sensitivity, and response time as crucial criteria. Thus, these findings can guide the fisheries industry in choosing effective and accurate formalin detection technology.</span>

Autonomous data sampling for high-resolution spatiotemporal fish biomass estimates

Many key ecological dynamics such as biomass distributions are only detectable on a fine spatiotemporal scale. Autonomous data collection with Unmanned Surface Vehicles (USV) creates new possibilities for cost efficient and high-resolution aquatic data sampling. However, the spatial coverage and sampling resolution remain uncertain due to the novelty of the technology. Further, there is no established method for analysing such fine-scale autocorrelated data without aggregation, potentially compromising data resolution. We here used a USV with an echosounder, a conductivity-temperature sensor and a flourometer to collect data from April–July 2019–2023 in a 60x80km area in the central Baltic Sea. The USV covered a total distance of 8000 nmi, over 42–81 days per year, with an average speed of 0.5 m/s. We combined the hydroacoustic data with publicly available oceanographic data from Copernicus Marine Service Information (CMSI) to describe seasonal distribution dynamics of a small pelagic fish community. Key oceanographic variables collected by the USV were correlated with CMSI estimates at daily/monthly resolution, respectively, to test for suitability to scale (Temperature 0.99/0.97; Salinity −0.77/−0.26; Chlorophyll-a 0.12/0.28). We investigated two approaches of Species Distribution Models (SDMs): generalized additive models (GAM) versus spatiotemporal generalized linear mixed effect models (GLMM). The GLMMs explained the observed data better than the GAMs (R2 0.31 and 0.20, respectively). The addition of environmental variables increased the explanatory capability of GAM and GLMM by 25 % and ∼ 3 %, respectively. Due to the high data resolution, we found significant amounts of positive autocorrelation (R: 0.05–0.30) across more than 50 sequential observations (>6 hours). However, we found that diel patterns in fish detection strongly affected the abundance estimates due to vertically migrating species hiding in the ‘acoustic dead zone’ near the seabed. Such dynamics could only be estimated and corrected for in predictions on the high-resolution data, complicating the trade-off between autocorrelation and high-resolution for SDMs. We compared estimates and effect sizes/directions in identical SDMs on 2x2km/month aggregated (i.e non-autocorrelated) observations and non-aggregated (i.e. autocorrelated) observations, and found relatively little difference in spatiotemporal estimates (r = 0.80). For the first time, we predicted the distribution of a small pelagic fish community at a high spatial resolution, in an area essential to breeding top predators, opening up for new applications in ecological studies locally and globally.

Fine spatio-temporal prediction of fishing time using big data

Overfishing, bycatch, and other anthropogenic threats may lead to the destruction of fragile habitats and substantial losses of marine life. Marine fishery resources can be protected by adjusting fishing intensity and establishing marine reserves. Currently, China adopts the closed fishing season management approach to protect traditional fishing grounds, where fine spatio-temporal prediction is essential to efficiently supervise the wide scope. Fishing vessel behaviors reflect fishers’ experience as well as the information provided by fish detection radar, while the fishery resource distribution is relevant to the marine environment. In this study, we identified fishing vessel behaviors (gillnets, trawls, purse seines, and abnormal behaviors) and qualitatively assessed and predicted fishing time of different fishing vessel behaviors to search for high intensity fishing operation areas by constructing a time-space prediction model. The model was based on big data of fishing vessel automatic identification systems and3 the marine environment, and was verified in the East China Sea. The prediction results generally corresponded with the distribution of traditional fishery resources in the East China Sea and the fishing efforts provided by the Global Fishing Watch. This model can provide an accurate and effective refined fishing vessel operation time prediction, and benefits fishing management and fishery resources protection.

Improving Inference Within Freshwater Community Studies: Accounting for Variable Detection Rates of Amphibians and Fish.

Research into freshwater communities often aims to link patterns of species distribution in ponds with underlying biotic factors. However, errors with species detection (e.g. false negatives) may underestimate distribution and bias assessments of community structure. Occupancy models that account for imperfect detection offer a solution to this problem. Here, we used three methods (call/visual encounter surveys, dip-netting and newt trapping) to survey amphibians and fish (potential amphibian predators) at 100 ponds in an urbanised landscape in Hungary over one breeding season. We estimated species detection probabilities for amphibians (all life stages combined) and fish using occupancy models to gain insight into amphibian-fish relationships and other survey-specific variables. We detected nine amphibian and 20 fish species. There were relatively low but variable estimated probabilities of detection for amphibians (mean: 0.320, 95% Bayesian credible interval: 0.142-0.598), with three species having detection rates < 0.1. Probabilities of detection peaked in the middle of the breeding season and increased with survey effort. Detection probabilities of five species were negatively associated with the detection of fish at a pond, while there were positive relationships between detection and emergent vegetation cover. We found no substantial differences in detection rates among the three survey methods. The probability of detecting fish was much higher than for amphibians (0.588, 0.503-0.717) but was lower at ponds with high emergent vegetation where amphibian detection was higher. Our results underscore the importance of accounting for the imperfect detection of both response organisms and potentially interacting species in aquatic community studies. We recommend applying multi-species occupancy models to enable inference for both common and rare species at ponds in landscapes subjected to human disturbances.

Research on Improved Lightweight Fish Detection Algorithm Based on Yolov8n

The fish detection algorithm is of great significance for obtaining aquaculture information, optimizing prey allocation in aquaculture, and improving the growth efficiency and survival rate of fish. To address the challenges of high complexity, large computational load, and limited equipment resources in existing fish target detection processes, a lightweight fish detection and recognition method based on the Yolov8 network, called the CUIB-YOLO algorithm, is proposed. This method introduces a secondary innovative C2f-UIB module to replace the original C2f module in the Yolov8 neck network, effectively reducing the model’s parameter count. Additionally, the EMA mechanism is incorporated into the neck network to enhance the feature fusion process. Through optimized design, the Params and FLOPs of the CUIB-YOLO algorithm model are reduced to 2.5 M and 7.5 G, respectively, which represent reductions of 15.7% and 7.5% compared to the original YOLOv8n model. The mAP @ 0.5–0.95/% value reaches 76.4%, which is nearly identical to that of the Yolov8n model. Experimental results demonstrate that compared with current mainstream target detection and recognition algorithms, the proposed model reduces computational load without compromising detection accuracy, achieves model lightweighting, improves inference speed, and enhances the algorithm’s real-time performance.

REAL-TIME DETECTION AND CLASSIFICATION OF FISH IN UNDERWATER ENVIRONMENT USING YOLOV5: A COMPARATIVE STUDY OF DEEP LEARNING ARCHITECTURES

This article explores techniques for the detection and classification of fish as an integral part of underwater environmental monitoring systems. Employing an innovative approach, the study focuses on developing real-time methods for high-precision fish detection and classification. The implementation of cutting-edge technologies, such as YOLO (You Only Look Once) V5, forms the basis for an efficient and responsive system. The study also evaluates various approaches in the context of deep learning to compare the performance and accuracy of fish detection and classification. The results of this research are expected to contribute to the development of more advanced and effective aquatic monitoring systems for understanding underwater ecosystems and conservation efforts.

Method validation for the detection of formaldehydein fish using ultra performance liquid chromatography

Formaldehyde, the simplest aldehyde, tops the list of illegal and harmful food contaminants. Recently, formaldehyde has been used indiscriminately to preserve the freshness and quality of fish. Numerous incidents of formalin (a solution of formaldehyde) misuse during the storage and transportation of fish have been reported in several parts of India. In this context, it is pertinent to study formalin contamination in fish and shellfish. Protocol for detection of free and bound formaldehyde by ultra-performance liquid chromatography (UPLC) was attempted in the present study. Formaldehyde peak was observed in the expected retentiontime and standardisation of the protocol for formaldehyde determination was then carried out. UPLC method was developed and validated for specificity, repeatability, accuracy, and linearity. The average correlation coefficient (R2) was 0.99. The recovery rates of free and bound formaldehyde were 90.83 to 116.43% and 88.76 to 132.98%, respectively. The limit of detection (LOD) was found to be 0.5 ppm. The study investigated free and bound formaldehyde levels in commonly traded fish species across five major fish markets inTamil Nadu, India. Sphyraena barracuda had the highest recorded concentrations of free and bound formaldehyde at 9.8 and 6.45 mg kg-1, respectively, while Lethrinus lentjan exhibited the lowest concentrations at 0.8 mg kg-1 for free formaldehyde and 0.5 mg kg-1 for bound formaldehyde. Keywords: Formalin abuse, Limit of detection, Method validation, Retention time UPLC

Survey on machine vision-based intelligent water quality monitoring techniques in water treatment plant: Fish activity behavior recognition-based schemes and applications

Abstract Water is a vital resource essential to the survival and development of all creatures. With the rapid growth of industry and agriculture, people face a severe threat of ecological destruction and environmental pollution while living earthly lives. Water pollution, in particular, harms people’s health the most. As a result, water supply security has become a top priority. As a critical point in water supply safety, monitoring water quality effectively and forecasting sudden water contamination on time has become a research hotspot worldwide. With the rapid development and wide applications of artificial intelligence and computer vision technologies, biological activity identification-based intelligent water quality monitoring methods have drawn widespread attention. They were taking fish activities as the water-quality indicator has gained extensive attention by introducing advanced computer vision and artificial intelligence technologies with low cost and ease of carrying. This article comprehensively reviews recent progress in the research and applications of machine vision-based intelligent water quality monitoring and early warning techniques based on fish activity behavior recognition. In detail, it addresses water quality-oriented fish detection and tracking, activity recognition, and abnormal behavior recognition-based intelligent water quality monitoring. It analyzes and compares the performance and their favorite application conditions. Finally, it summarizes and discusses the difficulties and hotspots of water quality monitoring based on the fish’s abnormal behavior recognition and their future development trends.

FishDet-YOLO: Enhanced Underwater Fish Detection with Richer Gradient Flow and Long-Range Dependency Capture through Mamba-C2f

The fish detection task is an essential component of marine exploration, which helps scientists monitor fish population numbers and diversity and understand changes in fish behavior and habitat. It also plays a significant role in assessing the health of marine ecosystems, formulating conservation measures, and maintaining biodiversity. However, there are two main issues with current fish detection algorithms. First, the lighting conditions underwater are significantly different from those on land. In addition, light scattering and absorption in water trigger uneven illumination, color distortion, and reduced contrast in images. The accuracy of detection algorithms can be affected by these lighting variations. Second, the wide variation of fish species in shape, color, and size brings about some challenges. As some fish have complex textures or camouflage features, it is difficult to differentiate them using current detection algorithms. To address these issues, we propose a fish detection algorithm—FishDet-YOLO—through improvement in the YOLOv8 algorithm. To tackle the complexities of underwater environments, we design an Underwater Enhancement Module network (UEM) that can be jointly trained with YOLO. The UEM enhances the details of underwater images via end-to-end training with YOLO. To address the diversity of fish species, we leverage the Mamba model’s capability for long-distance dependencies without increasing computational complexity and integrate it with the C2f from YOLOv8 to create the Mamba-C2f. Through this design, the adaptability in handling complex fish detection tasks is improved. In addition, the RUOD and DUO public datasets are used to train and evaluate FishDet-YOLO. FishDet-YOLO achieves mAP scores of 89.5% and 88.8% on the test sets of RUOD and DUO, respectively, marking an improvement of 8% and 8.2% over YOLOv8. It also surpasses recent state-of-the-art general object detection and underwater fish detection algorithms.

Molecular Cloning, Prokaryotic Expression, and Immunological Characterization of β-Enolase from Grass Carp (Ctenopharyngodon idella).

β-Enolase is a cross-allergen commonly found in fungi, plants, and aquatic products. Although studies on the allergenicity of fish enolase have been reported in recent years, they are still limited to a few species of marine fish. Therefore, the detection of freshwater fish in the food industry requires more studies of the molecular characterization as well as the allergenicity of enolase. In this study, the nucleotide sequence of β-enolase from grass carp was obtained by molecular cloning technology. Structural domain analysis showed that it contained the characteristic structural domains of the enolase superfamily, and homology analysis indicated that enolases are highly conserved evolutionarily. Recombinant β-enolase was obtained by prokaryotic expression, and its allergenicity was assessed by β-enolase-sensitized mice, which confirmed the ability of β-enolase to trigger an allergic response and cause a rise in Th1 and Th2 immune responses in mice. These results suggest that β-enolase could be used as a characterizing substance for the detection of fish allergens in the food industry as well as the preparation of drugs for allergy-related studies.

Rapid transformation of nanobodies affinity based on AlphaFold2's high-accuracy predictions and interaction analysis for enrofloxacin detection in coastal fish

High-affinity antibodies are crucial in biosensors, disease diagnostics, therapeutic drug development, and immunological analysis, making the enhancement of antibody affinity a key research focus within the field. Computer-aided design is recognized as a time-saving and labor-efficient method for nanobodies in vitro affinity maturation. Compared to experimental mutagenesis techniques, it is advantageous due to the elimination of the need for laborious library construction and screening processes. However, these approaches are constrained by structural prediction since inaccuracy in structure could readily result in maturation failures. Herein, a novel nanobodies modification method for in vitro affinity maturation, utilizing the high accuracy prediction of AlphaFold2, was employed to rapidly transform a low affinity nanobody against enrofloxacin (ENR) into one with high affinity. The molecular docking results revealed a 1.5- to 2.5-fold increase in the number of noncovalent interactions of modified nanobodies, accompanied by a reduction in binding free energy ranging from 14.1 to 62.6%. The evaluation results from ELISA and BLI indicated that the affinity of the modified nanobodies had been enhanced by 6.2–91.6 times compared to the template nanobody. Furthermore, the modified nanobodies were employed for the detection of ENR-spiked coastal fish samples. In summary, this research proposed a nanobodies modification method from a new perspective, endowing its great application potential in biosensors, food safety, and environmental monitoring.