Conference PaperPDF Available



Abstract and Figures

Fish breeding is a promising branch of farming, so the creation of tools for automation of this area is quite relevant. Feeding on fish farms is the main component of the successful functioning of such businesses. However, this process requires an in-depth preparation, as each species of fish has a different food culture, as well as various behaviours during nutrition. Moreover, in the method of feeding fish, farmers must take into account the age, size of the fish, and other characteristics. This paper contains information on the creation of a Preference testing by images processing is considered as the most effective tool that can be used to determine the sensory behaviour of an animal, which can record the eating behaviour of fish and determine the degree of their hunger, and, finally, to feed them. Moreover, small fish are shyer, which provokes their malnutrition. A smart feeding system can solve the issue of uniform the distribution of food for all fishes.
Content may be subject to copyright.
Natarajan Meghanathan et al. (Eds) : DaKM, SIPP, CCSIT, NCWMC - 2018
pp. 85–97, 2018. © CS & IT-CSCP 2018 DOI : 10.5121/csit.2018.81506
Mohammed M. Alammar and Ali Al-Ataby.
Department of Electrical Engineering and Electronics,
University of Liverpool, Liverpool, United Kingdom
Fish breeding is a promising branch of farming, so the creation of tools for automation of this
area is quite relevant. Feeding on fish farms is the main component of the successful functioning
of such businesses. However, this process requires an in-depth preparation, as each species of
fish has a different food culture, as well as various behaviours during nutrition. Moreover, in
the method of feeding fish, farmers must take into account the age, size of the fish, and other
characteristics. This paper contains information on the creation of a Preference testing by
images processing is considered as the most effective tool that can be used to determine the
sensory behaviour of an animal, which can record the eating behaviour of fish and determine
the degree of their hunger, and, finally, to feed them. Moreover, small fish are shyer, which
provokes their malnutrition. A smart feeding system can solve the issue of uniform the
distribution of food for all fishes.
Fish Feeding, Preference testing, Fish Farming, Smart Feeding System, Methods of Fish
Fish farming is perspective business, which grows rapidly as it is shown in the (Fig.1). Indeed,
fish feeding is one of the crucial forms of intensification of the fish-farming process [1]. Correct
fish feeding in farming allows applying more dense plantings, and, thereby, increasing the fish
productivity of ponds.
Figure 1. World fish production from 1950 to 2014 [2]
86 Computer Science & Information Technology (CS & IT)
Currently, fisheries around the world try to use various artificial food additives, which include all
the substances necessary for the healthy growth and development of fish [1]. Feeding the fish is
based on natural food, which the fish can usually find in natural water reservoirs, but on farms,
this food contains more vitamins and other nutrients (Table 1) [3].
Additionally, aquaculture is the cultivation of fish and crayfishes in the closed-cycle systems [4].
Like any business or occupation, aquaculture can face many risks and challenges. Aquaculture
makes it possible to grow living organisms in small volumes in conditions as close as possible to
natural ones. The task of any aqua-farmer is to create such conditions in which the fish can
behave comfortable and grow fast. Indeed, such growth is possible only in conditions, which are
as close to natural ones as possible, and include rational feeding of all species of fish, regardless
of their size and activity. More often, the natural habitat of a particular type of fish is in a much
worse form, including drying up and salinization of water bodies, and pollution with industrial
wastes. All this affects the natural habitat of fish. Finally, the task of the farmer is to create such
conditions that the fish would feel comfortable for reproduction.
Table 1. The types of food of fish on farms [3]
An intelligent feeding system is a quite simple and low cost which will motivate the fish farmers
to acquire it. Moreover, the system can reports will convey information about the number and size
of fish and their behaviour. Indicators will contain information about such categories of fish as
small, medium and large ones. Based on the analysis of the size of fish and their behaviour, the
farmer can draw conclusions about the correlation between these two indicators. However, an
intelligent feeding system can reduce the amount of used feed by 20% [5]. However, a false
positive rate, which detects hungry fish without reasons, can feed the fished when they are full.
The percentage of damage from such false feeding is not fully researched. This system will
provide the estimation of the preference fish feeding by using efficient ways can observe fish
behaviour and response with feeding which is significant for the farmers to improve fish
production in a short period of time and at low cost.
2. L
Fish farming has gained traction in recent years due to depleting stocks in the ocean that have
forced people to rear marine life in a domesticated environment. However, feeding has become a
challenge since vast amounts of food are wasted that may lead to water toxicity. In any case,
manual feeding is not efficiently leading to high operational costs. Development of an automatic
feeding system is highly advised since it enables a farmer to automate the process thus enhancing
efficiency. The Sustainable Aquaculture Feed System is a useful digital device that can identify
Computer Science & Information Technology (CS & IT) 87
fish species, sex, and count, which enables the farmer to develop a feeding program. Broadly, the
automation of fish feeding systems is bound to improve efficiency in aquaculture farms.
2.1 Review of Automatic Fish Feeding Techniques
Fish farming is a multi-billion industry that has evolved to incorporate technology into the
management that includes smart feeding systems. Towards this end, a comparison of several
feeding systems is vital to establish the operating mechanisms of the structures under review. To
begin with, the Sustainable Aquaculture Feed System has been equipped with a vision sensor
machine that is used to estimate the amount of feed required by the livestock being reared, which
prevents waste of nutrients availed to the fish [6]. Notably, the system can count the number of
fish that is critical in determining the amount of food required by the fish. In relation, the
technology can be used to deduce the size of the fish, which is essential in formulating the
feeding program to ensure that the amount released is sufficient. Even further, the approach can
detect the gender of the marine species since they require different nutritional amounts.
Finally, the model can be used to identify the type of fish in the farm, which is used to inform the
feeding program. The system has employed the bio-scanner to undertake the requirements as
mentioned above with considerable success. The SAFS system has several components that
include both hardware and software processing elements, which include a camera, Bluetooth
receiver, and input devices. The hardware part consists of a camera, sensors, and the feeder
system [6]. The incorporation of a graphical user interface is essential since it provides data
analysis structure that is used to study patterns in the tanks [7]. The use of Bluetooth is designed
to ensure information is relayed to the required area electronically, which allows that graphic
images can be shared within system components. The inclusion of a timer is critical to the success
of the feeding system since it ensures the fish are fed at the appropriate time with the video
having three frames image per second.
On the other hand, the development of an automatic feeder system has led to the creation of a
smart system, which is controlled using artificial intelligence. The idea is to monitor the feeding
process continuously utilising the interface. The device uses the Global Standard for Mobile
Communication, which enables the firm to track the progress of the feeding program remotely
[8]. Ideally, the invention is applied to issue commands to the feeding program in real-time.
Consequently, the system can operate with minimal human intervention resulting in a fully
automated product. The inclusion of a central processing unit is essential since it receives all
inputs in the system that is processed before being used to issue commands in the smart feeding
program. The system has a temperature sensor, which is used to analyse the conditions in the
water. It is imperative to state that water heat is critical in aquaculture since it influences the
deterioration of either the feed, which might affect positively or negatively the nutritional content.
The inclusion of a camera is designed to ensure digital images of the fish stock can be monitored
using the structure [9]. The camera enables the farmer to identify the number of fish, sex, size,
and number, which inform the amount and type of feed to be released into the farm. In fact, the
system has an 80% accuracy, which is good considering that the smart feeding industry is
relatively nascent. Being able to deduce the physical characteristics of the fish enables the farmer
to release the correct amount of feed, which reduces wastage, especially in controlled
aquaculture. Time management is another vital element that has been considered in the design
since the fish stock must be fed at appropriate times to ensure optimal nutritional value is derived
from the feeds.
88 Computer Science & Information Technology (CS & IT)
Table 2. Comparison of Different Fish Feeding System
The use of Automatic Computer Vision Systems for Aquatic Research is an efficient system that
has been developed to facilitate feeding of fish. The structure addresses several critical issues,
which include the incorporation of a sizing mechanism that can be used to inform the amount of
feed to be released into the farm. Further, the identification of fish species, specifically the zebra
fish is another critical component of the system that determines the type of feeds to be used in
feeding the marine animal. Again, the automated system enables the farmer to analyse the
behaviour of larval fish in the farm, which can deduce trends and patterns that can be used to
develop a feeding program. The system is equipped with a camera and computer processing
facility that can produce information [10].
In summation, depleting fish stocks in the ocean have led to the development of aquaculture
farms that require substantial amounts of feeds. The creation of fish feeding systems been
automated to ensure the process is efficient. Notably, manual feeding is wasteful and unable to
detect the nutritional needs of marine animals rightly. The Solar Powered Automatic Shrimp
Feeding System has integrated several components that ensure shrimps are fed at specific
intervals. The structure uses solar energy to release food into the tank, which improves energy
efficiency. To conclude, the automation of a feeding system in fish farms will enhance the
productivity of the products.
2.2 Challenges and Opportunities
The principal risks of aqua farming are fish diseases, technical malfunctions in the work of
equipment for feeding fish, saturation of water with oxygen, substandard feed, and others [12].
Additionally, usually, the feeding of fish is between 50% and 80 % of the overhead costs of a fish
farm [12]. Nutrition is a manual task, so it is an immeasurable and inaccurate method with such
results as overfeeding and underfeeding of fish. Overfeeding means that the majority of food goes
to waste, and it infringes the financial part of aqua farming, the surrounding the farm marine
environment, and the health of the fish. On the other hand, the lack of feeding leads to famine and
the gradual dying of fish. Moreover, some challenges may be faced with the course of this
project. These challenges may include:
1. Fish Size: Fish of various sizes take food differently. Small fish are more passive in the fight
for food because of their size, so they get insufficient food. Indeed, many fish are shy, and
they are at a far distance from the bold fish, which get food together. Shy fish receive an
inadequate amount of food, therefore, improving the method of feeding fish is a critical point
in the successful operation of aqua farms. The development of an automatic intelligent
feeding system is significant for solving these problems and distributing the right amount of
feed using sensors that measure the appetite of fish.
Nutrition and feeding have a significant effect on the health of fish. It affects their behaviour
and response to different environmental conditions. Since fish of different size have different
Market Sizing Feeding
Automatic Computer Vision Systems for Aquatic
Yes Yes Yes
Sustainable Aquaculture Feed System [6] Yes Yes Yes Yes
The Automatic Feeder System [8]
Solar Powered Automatic Shrimp Feeding System [11] Yes
Computer Science & Information Technology (CS & IT) 89
feeding patterns, it is prudent to learn their behaviour which helps the farmer to prepare an
effective fishing schedule. According to Lall and Tibbetts, fish behave differently depending
on the feeding habit, feeding method as well as the frequency of feeding [13]. The proposed
system will be able to analyse the food preference of a shoal for optimal growth.
2. Fish Behaviours: The behaviour of fish associated with nutrition depends on their type, size,
and sometimes on their sex [1]. Fishes usually eat other fishes or plants. Even representatives
of fish species, which do not belong to predators, eat small fish when their size is equal to the
preferred food. Indeed, adult fish mostly eat caviar and fry if they find it in the nearest area.
Moreover, fishes use to dig in the ground. Some fish directly get their own food in the
uppermost layer of soil, while others sift the ground through the gills and during this process
sometimes absorb reasonably large pieces of soil.
According to Lovell, of all the spectrum of behavioural reactions manifested by fish, the main
one is the behaviour associated with nutrition [1]. Food behaviour is a complicated process of
a successive change of individual behavioural phases and acts from the moment of obtaining
information about the presence in the environment of food objects before deciding whether to
seize or reject them [1]. The first phase of the nutrition behaviour of fish is the rest, which is
such state of fish when it does not react to the external food stimuli. It is common for the
majority of species, and it happens due to various causes, including illness, the closeness of
spawning period, wintering, etc. The second phase is the readiness for obtaining the signal on
food availability. The third phase is an obtaining signal on food availability. In the process of
scanning the water reservoir, the fish eventually discovers the signal emanating from the food
object. In the process, all sensory systems of fish are involved, so the signs received can be
variable in nature and have different intensity and direction. The next phase is the search and
detection of the source of food. Therefore, among the whole spectrum of available signals,
the fish chooses one. After the food signal is selected, fish start to search for the source. The
last phase is the determination of the suitability of food.
Moreover, Abdallah and Elmessery have stated that some fish used to eat only in open spaces
[14]. Others, which are shyer, hide in the clefts anticipating the best moment to swim out.
Cereal fishes spend a very long time feeding to satisfy their nutritional needs, while predatory
species like eels are not eaten every day. These varied ways of feeding become apparent in
the artificial ponds and require attention. Otherwise, the fish will not survive. Indeed, small
fish may not receive enough feed because they have shy behaviour patterns. According to
Abdallah and Elmessery, small fish may not appear near the sensors of hunger control, so
they may remain hungry, and would have to eat a minimum amount of snacks after huge
fishes [14]. Moreover, AlZubi thinks that this intelligent system of feeding fish requires the
significant campaign of advertisement and popularisation, since the majority of farmers are
not accustomed to the idea of using technology in their production process [5]. In the
beginning, many of them were resistant to adopting the application, but education and
training, as well as a rental system, allowed them to spend less money and give them the real
evidence [5].
3. Fish Tracking: According to Al-Jubouri, the current fish tracking methods require the tagging
of an individual fish which is quite challenging [10]. This calls for the need of advanced
system whereby the non-contact method of recognising a particular free moving fish has been
developed. The system does not only reduce the time for tagging process but also offer a real-
time recognition technique. The computer-aided tool in its different models provides a
successful solution for analysing the behaviour of fish, their feeding habit, and size. Studies
to find out whether fish have the capacity painful stimuli and associated discomfort have been
faced with a challenge of ethical restriction. Larval zebra has been used instead since their
responses are similar to those of the adults. It is therefore advisable to consider the ethical
90 Computer Science & Information Technology (CS & IT)
suitability of a given system that affects the behaviour of fish or any other animal being used
in the study.
3. A
The system should accomplish the task in three stages. The first stage in the algorithm is the
object detection where the subject under the study is identified. The object is then tracked and
monitored closely where the activities are recorded. For this case, a fish will be identified and
monitored. The collected data is processed and a conclusion drawn from the analysis. Three
stimulation techniques can be used in the study; they include thermal, electrical and chemical
stimulation. Electrical stimulation is the most opposed technique among the three since it inflicts
pain on the object under study. It is therefore used in a limited number of is also
associated with unpredictable behavioural reactions. It is also not easy to capture the movement
of small fish with this method due to the transparency in their body [10].
3.1 The Aims of This Study
In this study, to achieve the aims of study the system smart fish feeding. This will encompass:
1. In order to precisely identify and measure the amount of feed for each fish
2. Minimising the impact of traditional feeding mechanisms
3. Testing the fish preference for type of feed
4. Minimising the feed waste and maximising the conversion rate of food products
3.2 Objectives
The Objectives of the project can be defined as:
1. Optimising diet for fish
2. Developing a new way to keep the fish tank clean
3. Developing new ways of modulating data onto a smart fish feeding system
4. Determining the possibilities of feeding fish system recreation using digital data
4. E
4.1 Design considerations
This project will provide the estimation of the preference fish feeding by using ways to observe
fish behaviour and response to the type of feed. This project require tank is separated into two
areas; Living Area and Feeding Area which is the diet area. Living Area that is comfortable to
fish contains gravel and plants, however, Feeding Area is less comfortable because it is less
natural and bare to fish (Fig.2). The design shows the recirculation system which consists of a
water filter and pump that is used for filtration and measure the amount of waste of feed.
Furthermore, the sensors may not always fully reflect the state of the fish. Many water parameters
should be measured: pH, the temperature, salinity, dissolved oxygen, ammonium, the
transparency, the suspended solids, nitrates, the total nitrogen or match soluble reagent, among
others [15]. The smart fish feeding architecture comprises of a few components, which includes
the hardware design, and the software algorithm with the database of the fish.
Computer Science & Information Technology (CS & IT) 91
Figure 2. The smart fish feeding system setup
Modern methods of fish feeding include an intelligent feeding system based on fish behaviour
and extend to speed respond fishes toward one kind of feed to minimise the impact of traditional
feeding mechanisms. The proposed mechanism of nutrition interacts, recognises and reacts to the
activity of fish [5]. Such feeding system feeds fish at their request, regardless of the time of the
day. Figure 3, shows the block diagram of the system measure the Feeding Efficiency (FE), and
Specific Growth Rate (SGR, % body weight per day) which reflects fish respond development of
the given feed.
Figure 3: Block diagram of the proposed system
Where N number of tanks fish feeding which each tank feeding by one type of different feed. The
system connects to each tank that has the same volume of feed. The system could study the
preference testing for fish. Moreover, this design system is used to study the behaviour and
growth of fish toward a certain food. First of all, assume the fish does not have any knowledge
about the new fish feeding mechanism. In order to introduce the system to fish, the food is
dispensed based on schedule plan similar to traditional feeding mechanism called Adaptive
92 Computer Science & Information Technology (CS & IT)
Session. In this session, the fish feeding must be in the Feeding Area to learn fish the smart
feeding system. The smart fish feeding is running independently from the beginning of the
experiment in order to quantify fish behaviour and responses. When the fish show the high level
of learning the adaptive session, the system switch to the smart fish feeding system. A greater
learning factor is weighted using fish learning index when a more the system depends on the
behavioural feeder since actions from scheduled and behavioural feeders, which reflects fish
behaviour development during the adaptive session [5]. Further, the volume of feed consumed,
the number of time for fish seeking the food and the growth of fish is a great reflection to study
the extent of their response to a given feed through period time. The proposed system consists of
two sections: the Hardware and the Software.
4.2.1 Hardware:
The hardware of smart fish feeding system consists of a fish tank and three main components
listed as below:
1. Two Webcams: The webcam has a low cost for the farmer and equipped with various
devices to improve the quality of recording [5]. The first webcam is fixed over the
Feeding Area to take a top view of the fish activities. Moreover, the image snapshot from
webcam1 for counting and measure the length of fish. The second webcam is fixed in the
front side of the tank to take image snapshot to get measure the distance between the
object and the webcam1 and the estimated weight of an object by measure the Girth of
The Girth () of the fish was calculated using the following formulas:
Where a = semi-major axis length of an ellipse and b = semi-minor axis length of an
The Estimated Weight () of the fish modified [16]
Where is the length
(cm). Usually, a black and white image at greater depths is
better than a colour image. Black-and-white image has the advantage over colour views
of working in troubled waters with low transparency (Fig.4).
Figure 4. An example of shoot made by Logitech 720p webcam [6]
Computer Science & Information Technology (CS & IT) 93
2. Fish Feeder: The fish feeder is an automatic dispenser that is a horizontal cylindrical food
container with an adjustable gap at one end. The food container should be connected to a
stepper motor, which can control it by I/O multiplexer such as Arduino or Raspberry Pi
depend on the data from the webcams. Rotating the container 360 is dropped in the tank
a small portion of food.
3. Interface Circuit: The interface circuit consists the Arduino or Raspberry Pi and hardware
PC. The webcams and the fish feeder are connected to an interface circuit. The interface
circuit allows the software algorithm to control the fish feeder as a response to the fish
4.2.2 Software:
The software of the operation of such an intelligent feeding system is quite simple. After
determining the relative hunger of the fish through observe fish in the Feeding Area, the loaded
machine releases the configured number of feeds depends on the number of fish immediately and
sends real-time image snapshots directly to PC (Fig.5).
Figure 5. Flowchart of the proposed system algorithm
Based on the analysis information about the number and size of fish and their behaviour which
convey by the images. The PC can draw conclusions about the correlation of these indicators. The
image snapshots are the input source for the system which can analysis by using:
1. Graphical User Interface (GUI): The Graphical User Interface (GUI) the software
could develop using MATLAB which is a multi-paradigm numerical is computing
environment developed by Math Works. It allows matrix manipulators, plotting of
94 Computer Science & Information Technology (CS & IT)
functions and data, implementations of algorithms, a creation of UIs and interfacing
with text-based programs written in various languages such as C, C++, Java and
Python [17]. The developed software enables end-users to access:
Event system
Control of webcams
Control the feeding system
Counting fishes in the feeding area
Recording updating rate i.e. length and weight of fish
2. Object Detection: The object detection algorithm has a fundamental influence on the
performance of the counting and sizing systems. The system based on fish detection in
the Feeding Area.
3. Feeding Time (FT): The system measure the timing of fish feeding determines the
night or day mode of the system. Recording the timing of feeding is important to
calculate the volume of feed which dispenses. Further, calculation timing of feeding
helps to collect information about the time behaviour of fish during feeding as well as
the duration of each fish feeding.
4. The Vision System: The proposed methodology consists of nine distinct stages: image
acquisition, image pre-processing, image segmentation, feature extraction,
classification algorithm, and number, length, and girth of the fish estimation. Figure 6,
shows the block diagram of the proposed algorithm and are described as follows [18].
Figure 6. Proposed methodology of the vision system
i. Image Acquisition (Two webcams-Frames)
ii. Image Pre-Processing (Image Colour- Camera distance- Background extraction)
iii. Image Segmentation (Thresholding-Morphological operation- Watershed
iv. Feature Extraction (Extract size and shape feature)
v. Classification Algorithm (the sizing algorithm and counting algorithm)
vi. Estimated the Number of the Fish
vii. Estimated the Length of fish
viii. Estimated the Girth of fish
ix. Fish Length and Girth Monitoring and Counting Fish
Computer Science & Information Technology (CS & IT) 95
5. Feeding Decision: The feeding decision is based on the number of fish in the Feeding
Area which calculated by the vision system. If the number of fish value exceeded one
then feeding signal will be issued to the I/O multiplexer such as (Arduino or Raspberry
Pi controller) to turn the feeder on. Moreover, the feeding decision based on the
relative hung which is determining the overfeeding and underfeeding. This system
control variable volume of feed consumed that is calculated from the amount of waste
in the filtration system from each feeding.
Data Logging: During the specific long time, every update cycle is logged for further
analysis which is in night or day mode. The proposed system connect to the PC to
logging and analysis the data. The data shows specific growth ratio (SGR) and the
feeding efficiency (FE) during the period time. Moreover, the system exists the volume
of feed consumed which are accurate data to show the preference feed for fish because
the system dispenses the valid amount of stimulating feed by reducing the amount of
waste feed. The Feeding Efficiency (FE) and Specific Growth Rate (SGR, \% body
weight per day) were calculated using the following formulas:
Where FCR is Feed Conversion Ratio, M mean the mass of food consumed (g) and m
the increase in mass of animal produced (g)
Where W
and W
are the final and the initial body mass (g) respectively, and t is the total
number of days between the two measuring days [19].
5. C
aquaculture is a rapidly growing sphere of farming, which provokes the active
development of technology in this area. Fish feeding is an important component of fish farming,
so the inventions for improving feeding are relevant and requested. One such invention is the
smart fish feeding system, which can monitor food behaviour of fish using sensors. As an
illustration, one of such sensors is the webcam that can capture fish movements. A fully
automatic feeding system should be developed to understand fish's food behaviour in a dense
aquaculture tank. Such a system should have the ability to classify the activity of fish food intake
along with the continuous detection of the fish preference of excessive raw materials. Moreover,
the system should provide valuable information for controlling the feed in the food tank. The
algorithm of the system needs improvement constantly in order to provide information on how
fishes are active in numerical value and use detection of the boundaries of feeding for the more
in-depth understanding of the beginning of the feeding and its termination. This system should
also facilitate the creation of statistics on the increase of fish in order to develop computational
and quantitative approaches to a comprehensive understanding of the growth of fish.
96 Computer Science & Information Technology (CS & IT)
[1] Lovell, Tom, Nutrition and Feeding of Fish. Springer Science & Business Media, 2012.
[2] The Future of Agriculture, The Economist, 2016,
quarterly/2016-06-09/factory-fresh. Accessed December 15, 2017.
[3] BMA, Hasan, and B Guha, Optimization Of Feeding Efficiency In Semi-Intensive Farming System
For Sustainable And Cost Effective Production Of Penaeus Monodon Fabricius. Journal of Aqua-
culture Research& Development, vol 3, no. 6, 2012, OMICS Publishing Group, doi:10.4172/2155-
[4] Conti, Stephane G., Philippe Roux, Christian Fauvel, Benjamin D.Maurer and David A.Demer.
Acoustical Monitoring of Fish Density, Behavior, and Growth Rate in a Tank. Aquaculture, vol. 251,
no. 2-4, 2006, pp. 314-323. Elsevier BV, doi:10.1016/j.aquaculture.2005.06.018.
[5] AlZubi, Hamzah S., Waleed Al-Nuaimy, Jonathan Buckley and Iain Young. An Intelligent Behavior-
Based Fish Feeding System. 2016 13Th International Multi Conference on Systems, Signals &
Devices (SSD), 2016, IEEE,doi:10.1109/ssd.2016.7473754.
[6] Jer-Vui Lee, Joo-Ling Loo, Yea-Dat Chuah, Pek-Yee Tang, Yong-Chai Tan and Wei- Jian Goh. The
Use of Vision in a Sustainable Aquaculture Feeding System. Research Journal of Applied Sciences,
Engineering and Technology, vol 6, no. 19, 2013, pp. 3658-3669.
[7] Faizan Hasan MUSTAFA, Awangku Hassanal Bahar Pengiran BAGUL,Shigeharu SENOO, Rosita
SHAPAWI A Review of Smart Fish Farming Systems. Journal of Aquaculture Engineering and
Fisheries Research, vol. 2, no. 4, 2016.
[8] Md. Nasir Uddin, Mm Rashid, Mg Mostafa, Belayet H, Sm Salam, Na Nithe, Mw Rahman & A Aziz.
Development of Automatic Fish Feeder. Global Journal of Researchers in Engineering: A Mechanical
and Mechanics Engineering, vo. 16, no. 2, 2016.
[9] Zelda Dunn. Improved Feed Utilization in Cage Aquaculture by Use of Machine Vision. University of
Stellenbosch, 2008.
[10] Al-Jubouri, Quassay Salim, Automatic Computer Vision Systems for Aquatic. University of
Liverpool, Accessed September 2017.
[11] Dindo, T.Ani, Meryll, Grace F.Cueto, Niño, Jerome G.Diokno and Kimberly, Rose R.Perez, Solar
Powered Automatic Shrimp Feeding System. Journal of Multidisciplinary Research, vol.3, 2015, pp.
[12] Costa, C., A.Loy, S. Cataudella, D.Davis and M.Scardi Extracting Fish Size Using Dual Underwater
Cameras. Aquacultural Engineering, vol 35, no. 3, 2006, pp. 218-227. Elsevier BV.
[13] Lall, S P, and S M Tibbetts, Nutrition, feeding, and behavior of fish. The veterinary clinics of North
America. Exotic animal practice. U.S. National Library of Medicine, May 2009,
[14] Abdallah, S. E. and W. M. Elmessery, an Automatic Feeder with Two Different Control Systems for
Intensive Mirror Carp Production. Journal of Agricultural Engineering and Biotechnology, 2014, pp.
36-48. Bowen Publishing Company, doi: 10.18005/jaeb0203002.
[15] Miguel Garcia-Pineda, Sandra Sendra, Gins Lloret, Jaime Lloret. Monitoring and Control Sensor
System for Fish Feeding in Marine Fish Farms. September, 2011.
Computer Science & Information Technology (CS & IT) 97
[16] Mika, Kurkilahti, Magnus Appelberg, Trygve Hesthagen and Martti Rask, Effect of fish shape on
gillnet selectivity: A study with Fulton's condition factor, Journal of Fisheries Research. vol 54, no. 3,
2002, pp. 153-170. Elsevier BV.
[17] William J. Palm. Introduction of MATLAB for Engineers, University of Rhode Island, book Third
[18] Ibrahim Aliyu, Kolo Jonathan Gana, Aibnu Abiodun, James Agajo, Abdullahi M, Folorunso T ,and
Mutiu A, A Proposed Fish Counting using Digital Image Processing Technique. Journal of Science,
Technolology and Education (JOSTE), Federal University of Technology, 2017, pp. 36-48. Bowen
Publishing Company,ISSN:2277-0011.
[19] Sandie Millot, Jonatan Nilsson, Jan Erik Fosseidengen , Marie-Laure Begout and Tore Kristiansen,
Evaluation of self-feeders as a tool to study diet preferences in groups of Atlantic cod (Gadus
morhua), 2012.
Mohammed M. Alammar received MSc in Electrical Engineering from University of
Dayton, USA in 2016.He is a lecturer at King Khalid University, Abha, KSA. He is
currently pursuing Ph.D. degree with University of Liverpool, Liverpool, UK. His
research interest include Image Processing, Signal Processing and Embedded
... If the physical characteristics of fish are deduced, the farmers would be able to release a proper amount of food in the pond and prevent wastages (see Fig. 12.15). Time management is another crucial parameter that should be considered in the design of fish feeders as the fish stock must be fed at proper times, ensuring that optimal nutritional value is derived from the foods [101]. ...
... 15 The smart fish feeding system setup[101]. ...
The increasing trend of the global population has been doubled since the 1960s, and it is estimated to reach over 9.8 billion people by 2050 [1]. In this regard, nourishing the ever-increasing population and providing "Food Security" have been always global concerns. The statistics released by the Food and Agricultural Organization (FAO) in 2021 indicated that the number of people suffering from malnutrition has reached 690 million people [2]. To mitigate this issue, investing in the agriculture sector is more crucial than ever. Additionally, other global issues of climate change, water scarcity, and energy security have put excessive pressure on the agriculture and food production sectors, requiring more sustainable and eco-friendly agricultural operations [3,4]. One method to achieve this goal is to supply the energy demand of agricultural tasks using renewable energies that among them solar energy is the most abundant source with the highest adaptability with agricultural applications [5,6]. Although several pieces of research have studied the integration of conventional and modern agricultural operations with solar energy technologies such as solar-powered drying [7], solar-powered greenhouse cultivation [8,9], solar-powered irrigation and water pumping [10,11], solar desalination system [12,13], and solar-powered farm machinery [1,5], the investigation of some other emerging applications have been less considered. Therefore, this chapter presents a comprehensive study of some emerging applications of solar energy technology in agriculture, aquaculture, and food production 425 Solar Energy Advancements in Agriculture and Food Production Systems.
Full-text available
Dairy feeding causes significant water pollution. By controlling the proper amount of feed, reducing the waste to minimum will effectively reduce the problem of water contamination. In this project, a Sustainable Aquaculture Feed System (SAFS) has been designed and developed. It can automatically feed the fishes by estimating fishes' appetite through machine vision. The discussion includes design and optimization of the vision system using Labview as well as the integration of various components in the SAFS. With the developed algorithm, the system is able to detect the presence of fishes and count the number of fishes. The outcome is able to estimate and infer the fish appetite. Therefore, the feeding time can be planned ahead. In addition, the system includes a Graphical User Interface (GUI) for monitoring, display the feeding status and sensors reading such as pH, turbidity and temperature.
Full-text available
Feeding management provides the producer with an efficient tool to overcome limited feeding that may exhibit aggressive behavior during feeding due to limited feed availability resulting in carps that do not reach maximal growth. Overfeeding results in uneaten feed, poorer water quality, lower economic profit and additional environmental pollution. An automatic feeder was constructed and evaluated to provide predetermined amounts of food to four 9.5 m diameter tanks stocked with 7000 organisms. Three tanks worked as rearing tanks and the fourth as nursing tank. Growth methodology optimized carp production in tanks which are grouped in four. These big tanks require of a better food distribution so three points of the tank were selected: two aside the edge of tank wall and another close to the center. Each point or wireless controlled gate provided the necessary food giving more chance to the fishes to obtain their provisions without competition. The system uses a hopper capable of feeding the four tanks during a week and its precision was dependant on a weighting mechanism. Two ways were used to control food provisions in this work. An opened control system based on the ATM89c51 microcontroller controlled the exact dosing based on the tank requirements according to the carp cycle and the other closed loop control system was determined by the conditions of water temperature, fish age, body weight and the amount of oxygen consumed. The amount of oxygen consumed by carps was the best parameter knowing fish metabolism and growth that the feeder can rely on it controlling meals provisions. The results show minimal differences in growth (P<0.05) between treatments, important food saving of 25.337% (equivalent to 3495.5 kg), and lower water pollution (reduced water dissolved solids and ammonium components) compared with the first automatic feeder. Keywords: carp feeder, weighting mechanism, opened and closed loop control systems, oxygen consumption, and Intensive aquaculture systems.
Full-text available
Among other applications, self-feeding has been used to study food preferences in fish allowing them to choose between feeders with different food content. Preference tests assume that (i) trigger actuations are motivated by appetite, (ii) fish can learn which feeder contains which food and discriminate between feeders solely on the basis of their content, and (iii) in groups of fish, the triggering preferences is representative for the individuals of the group. We studied individual triggering behaviour in four groups of 14 Atlantic cod (length of 34 ± 2 cm, weight of 424 ± 102 g, mean ± SE, water temperature comprised between 7−8 ◦C) that were first given the choice between two self-feeders with identical content (Period 1 of 14 days) and subsequently with one feeder full and the other empty (Period 2 of 14 days) . In all four groups, one or two individuals performed the majority of the actuations, and in three groups the high triggering fish was a female high-ranked for size and growth rate. Cod displayed a preference for one of the two feeders despite their identical content. When the preferred feeder was emptied, the preference switched after one to eight days but both feeders were still actuated throughout the experiment. In conclusion, the assumption that actuation frequency reflects food preference and is representative for the individuals of the group may not be true, at least for Atlantic cod. If aiming at determining preferences representative for the whole population multiple representative fish should be kept isolated in separate tanks, with self-feeders containing each food option, on each tank.
Conference Paper
Aquaculture is a growing multi-billion pound industry facing many challenges. Traditional fish feeding mechanism in today's aquaculture farms stands behind a variety of challenges, including fish welfare, fish growth distribution, environmental effect especially in open ocean cage fish farms, and production cost efficiency. Adaptive smart fish feeder based on fish behaviours is proposed in this paper in order to minimize the effect of the traditional feeding mechanisms. The proposed feeding mechanism interacts, recognizes and responses to fish activities. The proposed smart fish feeder feeds fish based on their request regardless the time of the day. The smart fish feeding aims to minimize food waste and maximize the food conversion ratio (FCR). The proposed system is expected to cause uniform fish growth among individuals within the tank as the feeding depends on fish requests. Fish welfare is expected to be enhanced since there is no food competition and food waste is expected to be less making water good quality last for longer. This paper proposes hardware design of the smart feeder and smart software algorithm. Preliminary results will be discussed in this paper.
Nutrients must be provided in appropriate amounts and in forms that are biologically usable for optimum performance by the animal. Therefore it is as important to know the bioavailability of the nutrient as the dietary requirement. A respectable amount of data is available on digestibility of gross energy and crude protein in commercial ingredients used in fish feeds. There is, however, much less information on bioavailability of vitamins, minerals and amino acids from various natural and synthetic sources. In many cases, assumed availability values for nutrients are used to formulate fish feeds which are probably far from accurate. Examination of data presently available supports this contention.
Experiment with four different daily feeding frequencies, i.e., three (T1), four (T2), five (T3) and six times (T4) were conducted with supplementary feed (38% crude protein) in the earthen ponds (5000 m2) to determine the optimum feeding frequency for cost effective commercial production of Penaeus monodon Fabricius. Post larvae of black tiger shrimp (initial weight 0.02 ± 0.001 g) with stocking density of 20 m-2 were cultured for 110 days to evaluate the growth and production by studying different parameters of feed utilization efficiency as feed conversion ratio (FCR), protein efficiency ratio (PER), feed efficiency ratio (FER), production yield; and adequate growth levels as average body weight (ABW), weight gain (WG), specific growth rate (SGR), and survival of cultured shrimp. During production cycle, various water quality parameters of the ponds were found within normal range for aquaculture except for NH4 -N, NO3 -N and PO4 -P which were significantly lower in T3 (p<0.05, 0.01) and PO4 -P in T4 (p<0.05) series than T1. After harvest ABW, WG, SGR were significantly higher in T2, T3 and T4 ponds than T1 (p<0.05, 0.001). FCR, FER and PER followed the same pattern (p<0.05, 0.01, 0.001) in T2, T3 and T4 series than T1. Finally significantly higher (p<0.05) survival rate (82.5 ± 2.9%) and production yield (4562 ± 55.2 kg ha-1) were found in T3 than others, which indicated an additional support towards efficient feeding management and its outcome to T3. Maximized feed utilization and production efficiency with lesser wastage of feed was observed in T3 ponds than other series of ponds. Daily five-time feeding frequency (T3) earned significantly higher net profit of INR. 503233 ha-1 and return on investment (ROI) 69% compared to other treatments. Substantial improvement in growth and yield with higher profitability from ponds with daily five-time feeding frequency (T3) revealed it as the optimum feeding management for augmenting cost efficient production of P. monodon in a semi-intensive farming system.
The Eurasian perch, Perca fluviatilis (L.), was fished with Nordic multimesh gillnets, the mesh size combination of which was based on geometric series. From the test fishing data the fish girth was found to be linearly related to the third root of fish weight and therefore the fish girth could be estimated indirectly with Fulton’s condition factor K. Two different response surface models based on B-type gillnet selectivity model were fitted to the empirical data: one model with Fulton’s K and one without (simple model). The Fulton’s K-model had narrower Monte Carlo simulated 95% confidence interval than the simple model. Simulation studies showed that the traditional simple model without Fulton’s K does not have a linear relationship between mesh size and fish length, which violates the classical principle of Baranov’s theorem on geometric similarity of fish of different sizes. Therefore the simple model derived from one population is not applicable to another population if there are differences in fish condition. However, if as is the case for Nordic multimesh gillnets with a mesh size combination based on geometric series like Nordic multimesh gillnets, this error is minimal because adjacent mesh sizes cover each other and correct this error.
The definition of a remote system in the monitoring of fin fish growth rate and shape change relies on the development of the appropriate optical ranging system and the automation of data collection. Among the available technologies, a dual camera optical ranging system is presented as the most suited for fish size remote estimation. Images are collected via a submergible low-priced dual camera module connected to a portable waterproof PC equipped with two frame grabbers. Two images are synchronically collected and, with the system parameters (focal length of both cameras, distance between the two cameras, vertical and horizontal tilt and relative rotation of the two cameras), are used to collect data of fish size and shape. A Neural Network is built to correct the error of measurement. A geometric algorithm is developed to filter fish images and elliptic Fourier analysis of automatically collected fish outline coordinates is proposed as a tool of shape analysis. For all other fish orientations, landmarks (homologous points) are collected on fish outlines and evident structures: landmark configurations relative to each fish are then rotated using the system parameters. The proposed system can be used in the monitoring of sea-based fish farming facilities, especially those which are permanently submerged, reducing mortality and stress due to fish sampling and limiting divers intervention. By its application in the study of wild population, the system can be useful in the characterization of fish communities and population dynamics, supporting traditional visual census techniques with a tool which allows for continuous, remote and automatic data collection.