Automated Chicken Inspection
ARS agricultural engineer Yud-Ren Chen is developing a computer-directed
scanning system that could help speed inspection of the nearly 8 billion
chickens processed annually through federally inspected U.S. plants.
As chickens move down the processing line at speeds as high as 140 birds per
minute, four cameras click away, followed by near-infrared and visible light
scans of each bird.
Instantly, a computer decides whether a chicken has signs of defects or
disease. If not, the bird continues down the production line. Otherwise, the
computer directs the suspect carcass to a separate re-inspect line.
On the re-inspect line, birds get a closer examination by a human inspector
because the automated system spotted signs--such as reddish or purplish skin or
abnormally small body size--that suggested unwholesomeness.
That's the chicken plant of the near future, says Yud-Ren Chen, an
agricultural engineer with the Agricultural
Research Servicewho has led the group that designed and built the
prototype. He says the increasing popularity of poultry products has made
improved inspection even more important.
Chen's group will test their prototype this year at Tyson Food's poultry
processing plant in New Holland, Pennsylvania.
"Almost 8 billion chickens go through federally inspected plants
annually, compared to less than 3 billion 30 years ago," Chen says.
"If you are going to increase productivity without sacrificing the
accuracy of meat and poultry inspection, you have to use machine vision and
other automated sensors."
Developed over the past 7 years, the prototype consists of four spectral
cameras, a light probe, and a spectrophotometer--all linked to computers. When
the chickens, on hooks dangling from a moving chain, pass through a light beam,
the interruption triggers a fraction-of-a-second photo opportunity: One pair of
cameras takes photos of the chicken's front; the other pair, its back. One
camera of each pair uses a red filter; the other, a green one. This obtains
images of the bird's front and back in two colors.
"The same physical condition--involving surface color and
texture--shows up differently under different wavelength filters," says
Chen. "We use two wavelengths for comparison, to be sure we don't miss
Agricultural engineer Bosson Park uses a near-infrared and visible light probe
to scan a chicken's skin and underlying breast area tissue, to determine its
Chen's group developed computer software that compares the images at
different wavelengths, to determine if the bird is wholesome or not. Color
differences can be caused by improper bleeding during slaughter or by
blood-related diseases like septicemia. Skin textural differences can be caused
by tears, bruises, or tumors.
The cameras also detect body size. Chen explains that an abnormally small
chicken requires closer inspection because disease may have stunted the bird's
After a chicken passes the cameras, it crosses another light beam, this one
triggering a scan from about an inch away.
A light probe illuminates a portion of the chicken with both near-infrared
and visible light. The chicken absorbs some of the light, but any that is
reflected is analyzed by the spectrophotometer and computer using software
developed by Chen's team.
Differences between light shining on the bird and light reflection are due
to variations in external skin color and texture and to internal blood color
and tissue composition. In the prototype, the probe can analyze properties deep
beneath the chicken's skin, stopping only at the abdominal cavity.
A red light on the frame near the computer setup indicates rejection.
"The computer can also keep a record of the conditions of each bird on
the line, ready for the inspector's review," Chen says.
Chen's group has tested the system in a chicken plant in West Virginia. All
those tests used birds hung on a portable conveyor line brought to the
plant--alongside, but not on, a real production line like the one at the
Tyson's New Holland plant.
Leonard Payne, who manages the New Holland plant, has been looking forward
to this first production-line test of Chen's automated system.
The automated inspection system compares pictures of each bird as viewed
through a red filter and a green filter, to spot defective chickens. Here,
agricultural engineer Yud-Ren Chen places a green filter on one of the lenses
of the computerized, four-camera subsystem.
Payne says machine vision would benefit the industry mainly through more
consistency and accuracy. In most cases, wholesome birds differ obviously from
"Over 90 percent of the birds are unquestionably wholesome. Machine
vision could quickly pass these birds on, while identifying those that require
a second look to determine if a problem exists," says Payne.
"The federal inspectors and the assistants we provide are highly
trained. But machine vision can free them to focus on the relatively few birds
whose condition is not conclusive," he says. Payne sees machine vision, if
it works as expected, as a win-win situation for consumers, inspectors, and the
For Chen, an important part of the Tyson's test is to see how the system
stands up to the high humidity of a commer
cial production plant.
"We want to see how long the prototype lasts in this environment, how
much maintenance it'll need, and how accurate and consistent it is
online," says Chen. "The prototype has an average accuracy rate of
over 95 percent. We are continually improving this, and we achieved 100 percent
accuracy in a recent test comparing the system's conclusions with those of a
"To maintain accuracy," he says, "the system occasionally
needs retraining with special software. This adjusts the computer to recognize
the normal skin color of different chicken breeds or chickens fed different
rations, for example. Processing plant employees would do this retraining by
running self-learning software while flipping switches to show the system
chickens that are normal and chickens that are not."
Though this prototype can spot unwholesome birds, it can tell the reasons
for condemnation only in cases of septicemia or improper bleeding.
"But," notes Chen, "these two conditions account for over half
of the carcasses removed from the processing lines."
Chen aims to expand the system's capabilities, along with incorporating
advances in computer and sensor technology. He is also planning to test a new
probe that will explore the whole chicken--still without touching it--and take
color photographs of the abdominal cavity as well as the viscera. The color
images would also be analyzed by the computer.
Ultimately, Chen wants the automated system to quickly diagnose every
physical or biological condition that causes an inspector to remove chickens
from the processing line. It cannot spot bacterial contamination, he explains,
adding that many other scientists are busy developing tests for that.--By
Don Comis, Agricultural
Research Service Information Staff, 6303 Ivy Lane, Greenbelt, Maryland 20770,
phone (301) 344-2748.
Yud-Ren Chen is at the USDA-ARS
Sensing Laboratory, 10300 Baltimore Ave., Beltsville, MD 20705-2350; phone
(301) 504-8450, fax (301) 504-9466, e-mail
"Automated Chicken Inspection" was published in the May
1998 issue of Agricultural Research magazine. Click here to see this
issue's table of contents.