Chuanheng Sun

Near-infrared imaging to quantify the feeding behavior of fish in aquaculture

Delaunay Triangulation was applied to the extraction of behavioral characteristics.Support Vector Machine was used to classify the reflective frame.Serious reflection frames were removed and new data were fitted.The linear correlation coefficient between FIFFB and human expert can reach 0.945. In aquaculture, fish feeding behavior under culture conditions holds important information for the aquaculturist. In this study, near-infrared imaging was used to observe feeding processes of fish as a novel method for quantifying variations in fish feeding behavior. First, images of the fish feeding activity were collected using a near-infrared industrial camera installed at the top of the tank. A binary image of the fish was obtained following a series of steps such as image enhancement, background subtraction, and target extraction. Moreover, to eliminate the effects of splash and reflection on the result, a reflective frame classification and removal method based on the Support Vector Machine and Gray-Level Gradient Co-occurrence Matrix was proposed. Second, the centroid of the fish was calculated by the order moment, and then, the centroids were used as a vertex in Delaunay Triangulation. Finally, the flocking index of fish feeding behavior (FIFFB) was calculated to quantify the feeding behavior of a fish shoal according to the results of the Delaunay Triangulation, and the FIFFB values of the removed reflective frames were fitted by the Least Squares Polynomial Fitting method. The results show that variations in fish feeding behaviors can be accurately quantified and analyzed using the FIFFB values, for which the linear correlation coefficient versus expert manual scoring reached 0.945. This method provides an effective method to quantify fish behavior, which can be used to guide practice.

A traceability system for beef products based on radio frequency identification technology in China

Abstract Radio frequency identification (RFID) technology is an automatic identification and data capture (AIDC) technique and it is the representative technology for handling beef traceability. In this paper we offer a complete solution including: frequency, identifier information system, information system, data organisation, tag reclaimation and control technology, from three segments of the beef production process in China. From the farm to the slaughterhouse, electronic identification ear tag technology was used to identify the individual animal. From the slaughterhouse to the processing plant, gambrel identification was used to transfer the carcass information from the “old ear tag” to a “new ear tag”. Last, gambrel RFID was converted to UCC/EAN‐128 barcode labelling with a wireless electronic scale. The whole solution was evaluated successfully in Beijing.

Near infrared computer vision and neuro-fuzzy model-based feeding decision system for fish in aquaculture

Near infrared vision was used to quantify feeding behavior of fish.Fish feeding decision was realized using the neuro-fuzzy model.The method performance was evaluated by accuracy, fish growth and water quality.The proposed method can save feed costs and reduce water pollution. In aquaculture, the feeding efficiency of fish is of great significance for improving production and reducing costs. In recent years, automatic adjustments of the feeding amount based on the needs of the fish have become a developing trend. The purpose of this study was to achieve automatic feeding decision making based on the appetite of fish. In this study, a feeding control method based on near infrared computer vision and neuro-fuzzy model was proposed. The specific objectives of this study were as follows: (1) to develop an algorithm to extract an index that can describe and quantify the feeding behavior of fish in near infrared images, (2) to design an algorithm to realize feeding decision (continue or stop) during the feeding process, and (3) to evaluate the performance of the method. The specific implementation process of this study was as follows: (1) the quantitative index of feeding behavior (flocking level and snatching strength) was extracted by Delaunay Triangulation and image texture; (2) the adaptive network-based fuzzy inference system (ANFIS) was established based on fuzzy control rules and used to achieve automatically on-demand feeding; and (3) the performance of the method was evaluated by the specific growth rate, weight gain rate, feed conversion rate and water quality parameters. The results indicated that the feeding decision accuracy of the ANFIS model was 98%. In addition, compared with the feeding table, although this method did not present significant differences in promoting fish growth, the feed conversion rate (FCR) can be reduced by 10.77% and water pollution can also be reduced. This system provides an important contribution to realizing the real-time control of fish feeding processes and feeding decision on demand, and it lays a theoretical foundation for developing fine feeding equipment and guiding practice.

An adaptive image enhancement method for a recirculating aquaculture system

Due to the low and uneven illumination that is typical of a recirculating aquaculture system (RAS), visible and near infrared (NIR) images collected from RASs always have low brightness and contrast. To resolve this issue, this paper proposes an image enhancement method based on the Multi-Scale Retinex (MSR) algorithm and a greyscale nonlinear transformation. First, the images are processed using the MSR algorithm to eliminate the influence of low and uneven illumination. Then, the normalized incomplete Beta function is used to perform a greyscale nonlinear transformation. The function’s optimal parameters (α and β) are automatically selected by the particle swarm optimization (PSO) algorithm based on an image contrast measurement function. This adaptive image enhancement method is compared with other classic enhancement methods. The results show that the proposed method greatly improves the image contrast and highlights dark areas, which is helpful during further analysis of these images.

Fuzzy classification of orchard pest posture based on Zernike moments

Identification and count of orchard pests is very important in monitoring orchard pest population. The pests trapped by high-voltage grid show different postures and incomplete bodies, which increase the difficulty of image automated identification. Currently, most researches of pest image identification focus on feature extraction based on standard posture samples, without considering the influence from multi-pose of pests in natural scene. Consequently, the identification rates of these methods are low in practical orchard application. Using Dichocrocis punctiferalis (Guenee) as research object, this paper is directed towards a posture classification method for the orchard target pest identification. It aims at intensifying the performance of multi-pose pest identification system by utilizing Zernike moments as descriptors of shape characteristics. The input image is cropped automatically and further subjected to a number of preprocessing stages. The outcome of preprocessing stage is one processed image containing scaled and translated target pest. Then, the template number is determined according to the posture of target pest and the corresponding template parameters are obtained from the cluster centers by fuzzy C-mean clustering method. Experiment results show that the proposed shape feature is robust to changes caused by pest image shape rotation, translation, and/or scaling. And the highest accuracy of posture classification is 92.3% for orchard target pest Dichocrocis punctiferalis (Guenee) with multiple postures. It outperforms the method in reference [2] where the highest accuracy is 86.6%.

Correlation search between growth performance and flock activity in automated assessment of Pekin duck stocking density

Abstract In recent years, duck production has been changed from conventional free range and open water outdoors to confinement in birdhouses. And the concept of animal welfare in high stocking density begins to be accepted in China. The search of the relationship between growth performance and flock activity of Pekin ducks with different stocking densities using camera surveillance has great potential as an aid to improving flock management. The aim of this study was to determine the impact of stocking density on growth performance and activity of Pekin duck flocks. Furthermore, the correlation between growth performance and the activity of White Pekin ducks was investigated. Eventually, it will generate an automatic method for monitoring growth performance and stocking density evaluation of Pekin ducks. All ducks (24 days of age, n = 1200) were randomly allotted into 5 stocking density groups of 5 ducks/m2, 6 ducks/m2, 7 ducks/m2, 8 ducks/m2, 9 ducks/m2, with 6 replicates for each group. One group was selected for monitoring the activity of ducks using video recording system. The optical flow measures were extracted to describe the duck flock activity statistically using night video data. On the 24 and 42 days of age, sample collection was conducted for initial and final body weight (BW) measurements respectively. The results showed that the stocking density had significant effects on final BW and weight gain (P   0.05). The stocking density also had significant effects on mean and variance of optical flow produced by duck flock activity throughout the experimental duration. In the last, a significant relation was found between the mean of optical flow and final BW (r2 = +0.87). These results show that the proposed method has potential in automatic growth monitoring and stocking density management of Pekin ducks. It can cut out the bio-security risk and animal stress of having people actually visiting duck houses.