Fish processing using computer vision and robots

Arnarson, H.; Khodabandehloo, K

doi:10.1007/978-1-4615-2129-7_2

Cited by 5 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to a team of cooperating human workers, fish processing plants today are dominated by the use of relatively disconnected individual processes and inflexible process equipment, with minimal communication between processes and with human intervention to correct the out-feed of one process before the in-feed of the next. In addition, the majority of quality grading tasks remain manually based (Bondø et al, 2011;Arnarson and Khodabandehloo, 1993). Current policy chosen by the Norwegian Government states that automation of fish processing is the strategy preferred to increase the competitiveness of fish food products processed in Norway, by increasing product quality and reducing production costs (predominantly the costs of manual labour) through automation of fish food processing (Stortingsmelding: Whitepaper, 2013), the Norwegian Research Council (Hav21 2012) and the Norwegian Seafood Federation (FHL, 2012;FHL 2013).…”

Section: Description Of the Fish Processing Industrymentioning

confidence: 99%

A case study of obstacles and enablers for green innovation within the fish processing equipment industry

Bar

2015

Journal of Cleaner Production

View full text Add to dashboard Cite

Section: Description Of the Fish Processing Industrymentioning

confidence: 99%

A case study of obstacles and enablers for green innovation within the fish processing equipment industry

Bar

2015

Journal of Cleaner Production

View full text Add to dashboard Cite

“…Previous work on integration of intelligent sensing, robots and end‐effector technology for fish inspection and processing tasks has resulted in several solutions based on machine vision and robots (Arnarson and Khodabandehloo, 1993; White et al , 2006; Buckingham and Davey, 1995; Buckingham et al , 2001). Robofish 1 was developed as a prototype high‐speed, vision‐guided robot (two rotating arms) with an end‐effector (two‐fingered gripper) that is able to grasp slippery fish and accurately place them within a de‐heading machine (Buckingham and Davey, 1995).…”

Section: Introductionmentioning

confidence: 99%

An automated salmonid slaughter line using machine vision

Bondø

Mathiassen²,

Vebenstad³

et al. 2011

Industrial Robot: An International Journal

View full text Add to dashboard Cite

PurposeThe purpose of this paper is to describe a new slaughter line for industrial slaughtering of salmonid fish. Traditionally, slaughtering of farmed salmonids – salmon and rainbow trout – was done manually by bleed cutting with knives. Using the new slaughter line that includes 3D machine vision and a bleed‐cutting robot, slaughtering is almost completely automated – nominally requiring only one person to supervise the line and manually bleed cut the fish not handled by the robot.Design/methodology/approachThe design approach of the salmonid slaughter line focuses on using 3D machine vision and a bleed‐cutting robot with four biaxial pneumatic actuators to handle the slaughtering of pre‐anesthetized salmon and rainbow trout.FindingsUnder normal operating conditions, the slaughter line is capable of automatically slaughtering 85‐95 percent of all fish at an average feed rate of 30‐80 salmon/min, and the remaining 5‐15 percent are slaughtered manually. Several issues have been discovered, that should be addressed to improve the slaughter line.Originality/valueThis paper presents a new complete salmonid slaughter line that has reduced the need for manual labor in salmonid slaughtering plants.

show abstract

“…Some of the earliest research in fish length estimation using computer vision techniques was done in the early 1990s, Nielson et al (1991) [60] discussed the potential uses of an automated system for quality assurance using vision techniques, though they summarised that commercially available equipment to perform this with the right precision, was at the time unavailable. A paper in 1993 Arnarson et al described a prototype machine in which fish were passed under a video camera over a conveyor belt and were then sorted into bins, achieving 99% accuracy for flatfish and round fish [21].…”

Section: Related Workmentioning

confidence: 99%

Machine Learning for Tarakihi Fish Length Estimation in Aotearoa

Stanley¹

View full text Add to dashboard Cite

Current practices for monitoring the catch of deep sea fishing vessels is labour intensive requiring a person on vessel measuring individual fish lengths manually. Capturing videos of fish on-vessel instead allows the use of machine learning algorithms for tackling a computer vision based problem to automate the collection of morphological data of the observed fish. In this thesis we investigate essential methods required and develop a system that uses machine learning algorithms and computer vision techniques to calculate centimetre accurate lengths of singulated fish from video footage. The first stage in this process is the data acquisition, where we explore the use of both a fixed camera and a free camera (one that is held in hand) for gathering the video data from which we extract millimeter lengths. Lengths were gathered on-site to compare the lengths found from images at different orientations and translations. An analysis of the different camera positions and rotations found that a camera positioned above the object of interest and calibration pattern was able to achieve the most accurate lengths. Rotation was found to have an increasingly detrimental effect on predicted lengths as rotation, measured in radians, increased. Secondly, we use binary masks that are created both manually, and by using an automated approach, based on edge detection, for training a segmentation model to identify fish in images. We leverage shape features and interpretable ML classifier to analyse the features of contours from both inferred masks and those derived from an edge detection. In our analysis of these shape features, we identify a range of values for the feature ”circle deviation” which may be used to identify potential fish contours that did not pass the classification, and flag such contours for further training. We use contours, derived from the edge detection approach, that do pass the classification for creating a dataset of cropped images, to train a GAN from which a synthetic imagery set is created. Thirdly, we develop a method from extracting the lengths of fish from images by using a checkerboard pattern as a point of reference, to relate pixel lengths to millimeters. Only inferred contours with shape features within the range identified by our analysis are used in calculating lengths. This approach reduced the number of partially visible fish or false positive inferences from affecting our recorded lengths. Finally, the performance of models trained on real images, synthetic images, and a combination of the two are compared. A model that was trained on both real and synthetic images achieved an average for the absolute differences between true and predicted lengths of below one centimetre over 128 samples. Our results suggest that the use of synthetic data to assist in the creation of a robust training dataset is viable. However, this synthetic data works best when there is also real data available in the training set.

show abstract

Fish processing using computer vision and robots

Cited by 5 publications

References 9 publications

A case study of obstacles and enablers for green innovation within the fish processing equipment industry

A case study of obstacles and enablers for green innovation within the fish processing equipment industry

An automated salmonid slaughter line using machine vision

Machine Learning for Tarakihi Fish Length Estimation in Aotearoa

Contact Info

Product

Resources

About