Description
An apparatus for determining moisture content of an object
and a method thereof
The present invention relates to a system and a method for
determining moisture content of an object, particularly by
use of infrared imaging.
Moisture may be defined as a relatively small quantity of
diffused water. All materials contain at least a diminutive
volume of moisture as part of their molecular makeup.
However, the relative percentage of moisture in a substance
is dynamic.
Any given material will absorb moisture relative to ambient
temperature and humidity conditions through hygroscopic
action. Moisture content is a critical component of material
quality and its measurement is essentially a function of
quality control in most production and laboratory facilities.
In an entire spectrum of organizations ranging from
pharmaceutical manufacturers to food producers, moisture
content greatly influences the physical properties and
product quality at all stages of processing and final product
existence.
Weight, thermal expansion, amalgamation, and electrical
conductivity are some of the properties of a material that
are influenced by a change in the moisture content. A
processed material's moisture content will define, for
example, the shelf life of processed foods and provisions,
the reactivity of chemical compounds in inventory, or the
binding properties of bulk materials. As a result, the
ability to accurately identify moisture content levels during
processing procedures is paramount to the success of
countless commercial and scientific operations.

The measurement of moisture content using thermo gravimetric
analysis defines moisture as the loss of mass of a substance
when heated, by the process of water vaporisation. The
difference in mass is continually calculated and recorded by
a precision balance. Sample substance mass is measured before
and after the drying process for final moisture determination
on percentage basis. Methods for heating include oven drying,
microwave drying processes, halogen and infrared drying. The
disadvantages of the technique are its destructive nature and
the time consumed.
The measurement of moisture using absorption of radiation
involves irradiation of a moisture-bearing "sample with light
in the near infrared spectrum. The reflectance or absorbance
of the rays could be studied to analyze the moisture content
of the sample. This principle is used for calibration of the
meter, so that the moisture content may be displayed by
reading this reflectance electrically. The disadvantage of
the technique is' that the moisture content of the sample as a
whole cannot be determined; one only arrives at the moisture
content at a specific location on the sample. Thus, we do not
arrive at a holistic picture of the moisture content of a
sample.
It is an object of the present invention to provide a non-
local and non-destructive way of measuring moisture content.
The said object is achieved by an apparatus for determining
moisture content of an object according to claim 1 and by a
method of determining moisture content of an object according
to claim 9.
The underlying idea is to determine the moisture content of
an object using a histogram computed from an infrared camera
image of the object. The histogram shows a peak associated
with the object, whose position in said histogram is first
identified. The apparatus further maps the position of the
peak to a moisture value. By using a camera image as basis

for a quantitative measurement of the moisture content the
entire object is taken into account without the need to
modify the object.
In general it is known to use infrared cameras for the
inspection of homes that have water damage, moisture
intrusions or visible mould. However, the cameras are used
only to detect the presence or absence of moisture and cannot
be used to quantify moisture content as in with the current
invention.
In a preferred embodiment, the identifying means identifies
the peak in the histogram based on a known relation between
the temperatures of the object and surrounding of said
object. A relative knowledge of the temperature between the
object and its surrounding helps in identifying the peak
associated with the object in the histogram. For example, the
surrounding could include the atmosphere or a means that use
to hold the object in place in order to enable an image
capture.
In a further preferred embodiment, identifying means
identifies the peak in the histogram by comparing peaks in
the histogram associated with at least two reference objects
in the camera image. The at least two reference objects in
the camera has different known temperatures. The presence of
the peaks associated with the reference objects in the
histogram helps in identifying the peak associated with the
object much faster. More over by using specific reference
objects the influence of external environmental conditions
could be minimised.
In an alternative embodiment, the means for mapping, maps the
position of the peak to the moisture value based on a
function. The function is derived based on a pre-calibration
done for a specific type of object. The use of the function
helps in fast translation of the position of the peak to the
moisture value.

In an alternative embodiment, the means for mapping further
comprise a database with a plurality of data points relating
the position of peaks with moisture values from which the
moisture value of said object is interpolated. This is an
alternate approach to get the moisture value by means of a
database. The interpolation technique will enable to find the
approximate moisture content using nearby data points, even
if the exact mapping of the position of the peak and the
moisture value is unavailable in the database.
In an alternative embodiment, the means for mapping comprises
a neural network, said neural network trained based on the
pre-calibration. The data available during the calibration
phase will be harnessed by neural network to get trained.
This enables to indicate the moisture content instantaneously
to the user.
The present invention is further described hereinafter with
reference to illustrated embodiments shown in the
accompanying drawings, in which:
FIG 1 illustrates a block diagram of an apparatus for
determining moisture content of an object according to an
embodiment of the invention,
FIG 2 illustrates a histogram of a camera image of the object
captured by the infrared camera according to an embodiment of
the invention,
FIG 3 illustrates an arrangement to capture infrared camera
image according to an embodiment of the invention using at
least two reference objects,
FIG 4 illustrates a histogram of a camera image having the
peak associated with the object and at least two reference
objects, and

FIG 5 illustrates a flowchart describing the method of
determining of moisture contest of an object.
FIG 1 illustrates a block diagram of an apparatus for
determining moisture content of an object according to an
embodiment of the invention. The apparatus 100 is used for
determining the moisture content of an object 110. The
apparatus has an infrared camera 120 directed towards the
object 110 to capture a camera image. The idea here is to
measure the moisture content of a material by analyzing the
image of the object captured using an infrared camera. From
the captured image, a computing means 130, computes a
histogram for the camera image. Generally, since the object
is the only prominent object in the camera image, it will be
identified as a peak in the histogram. The apparatus further
comprises of an identifying means 140 to identify a peak in
the histogram corresponding to the object 110. A means 150
finally maps the position of the peak corresponding to the
object 110 in the histogram to a moisture value. The
apparatus further comprises an output device 160, for example
a display device or a loud speaker, which indicates or
communicates the moisture value to the user. The use of the
camera image for moisture detection enables the process to be
a non-destructive method. Hence the object under examination
can be returned to the production line at the end of the
measurement process. As there is no loss of object under
examination, more frequent inspections can be carried out.
Also the said technique does not alter the physical or
chemical nature of the object in any way.
FIG 2 illustrates a histogram 200 of a camera image of the
object captured by the infrared camera 120 according to an
embodiment of the invention. For a camera image, the
histograms show the image's overall exposure. Using 256
vertical bars to represent brightness levels from 0 to 255,
the leftmost bar is the darkest pixel level 0, and the
rightmost bar is the lightest level 255. The height of the
bars represents the total number of pixels at that brightness

level. The image of an infrared camera 120 is usually
displayed in pseudo color or gray scale. In a pseudo color
image, objects at different temperatures appear in different
colors. The assignment of colors to temperature is not
constant, but dynamic. The camera assigns different colors to
different temperatures. A similar process is also used for a
gray scale image. A gray scale image is composed exclusively
of shades of neutral gray, varying from black at the weakest
intensity to white at the strongest. On a numerical scale (8-
bit), the color black corresponds to 0, and the color white
corresponds to 255. For explanation, a gray scale images are
considered here. The proposed algorithm is equally applicable
to pseudo color images. The Y axis 210 represents the number
of pixels and the X axis 220 represents the gray scale
values. The peak 230 represents the histogram peak associated
with the object 110. The height of the peak 230 represents
the total number of pixels at that brightness level 240. Here
the identifying means 140 as discussed in FIG 1, identifies
the peak in the histogram based on a known relation between
the temperature of the object 110 and surrounding of said
object 110. The position of the peak is then mapped to a
moisture value, which can be indicated or communicated to the
user.
FIG 3 illustrates an arrangement 300 to capture infrared
camera image according to an embodiment of the invention
using at least two reference objects. The set up consists of
a small enclosure 310, with a pedestal 320 on which the
object 110, whose moisture need to be found will be placed.
The infrared camera 120 is affixed at the top of the
enclosure 310. A suitable lens is chosen for the infrared
camera 120 such that an object placed on the pedestal 320 is
clearly in focus. The object 110 is placed upon the pedestal
320 and enclosure 310 is closed so that the external
environment does not exert any influence upon the imaging
process. A heating element 330 is installed within the
enclosure 310 and is positioned such that it falls within the
field of view of the infrared camera 120. The electric

current running through the heating element 330 is controlled
such that the heating -element 330 is maintained at a constant
temperature. This temperature could be chosen to be higher
than the. temperature of any object within the field of view
of the camera, including the object 110.
The image of an infrared camera.120 is usually displayed in
pseudo color or gray, scale. The heating element 330 serves
the purpose of a color reference. Since it is always the
hottest object within the field of view of the camera 120, it
will always be assigned a color at the start of the color
band, which is reserved for the hottest object, e.g. the
color white in a gray scale image. Here the temperature of
the pedestal 320 is also known prior to doing the said image
capture. Hence the heating element 330 along with the
pedestal 320 acts as two reference objects.
FIG 4 illustrates a histogram 400 of a camera image having
the peaks associated- with the object and at least two
reference objects. Color of an object in an infrared image is
directly related to the temperature of the object. In a gray
scale image, hotter.objects appear white, and cooler objects
appear black. Furthermore, there are two possibilities: (a)
the object in question is at a higher temperature with
respect to the pedestal, and (b) the object in question is at
a lower temperature with respect to the pedestal. Here, the
heating element placed within the field of view of the camera
is maintained at a temperature higher than that of the
pedestal and the material in question. Thus, the peak 410 in
the histogram with the highest gray scale value (white / 255)
. will correspond to the heating element 330. Prior knowledge
of the temperatures of the pedestal 320 will allow us to
identify the peak corresponding to the said pedestal 320,
Instead of the pedestal 320 any other receiving means for the
object can be used.
In the current scenario, the object 110 is at a higher
temperature with respect to the pedestal 320. The peak 420

represents the associated peak of the pedestal 320. The
knowledge of the temperatures of the reference objects helps
to indentify the third peak 430 in the histogram
corresponding to the object 110. Thus the identifying means
140 identifies the peak in the histogram by comparing peaks
in the histogram associated with at least two reference
objects in the camera image. The reference objects generally
would have different known temperatures. The relative
position of the peak 430 with respect to the peaks 420 and
410 of the reference objects is mapped to a moisture value.
The said moisture value, could be indicated to the user
accordingly.
The mapping is done using a means 150 as shown in FIG 1. The
means 150 for mapping, maps the position of the peak to the
moisture value based on a function. The function is derived
based on a calibration done for a specific type of object.
The step of calibration involves measuring the moisture
content of the material using any one of the traditional
methods of moisture measurement.
The means 150 for mapping could even comprise a database with
a plurality of data points relating the position of peaks
with moisture values. These data points could be pre-
calibrated and associated with moisture values using some
traditional methods of moisture measurements. By knowing the
position of the peak of the object in the histogram, the
exact moisture value of said object can be interpolated using
the available data points.
The means 150 for mapping could also be a neural network. The
said neural network could also be trained based on the pre-
calibration as discussed above. A neural network consists of
a network of simple processing element called neurons. A
neural network can be used for function approximation that is
to learn the relationship between two quantities. A neural
network can be trained to learn a function that maps the
position of the peak of the object or the relative position

of the peak of the object from the peaks of the reference
objects, to a moisture value. The data available during the
calibration phase will be harnessed by neural network
training algorithms to train the neural network. At the end
of training, the neural network should output the moisture
content of the material in question.
The functional relationship between the temperature and the
moisture content of the object will depend on the material
the object is made of. Therefore a pre-calibration for
different types of material will increase the accuracy of the
measurement of the moisture content. The .function and the
database mentioned above can represent this functional
relationship for different types of materials at the same
time.
FIG 5 illustrates a flowchart describing the method of
detection of moisture content of an object. At step 505, an
infrared camera captures the image of the object, whose
moisture needs to be determined. At step 510, a histogram for
the camera image is computed. At step 515, the peak
associated with the object in the histogram is identified.
This can be done in two ways. First, based on a known
relation between the temperatures of the object and
surrounding of said object as shown in step 520. The
surrounding could even include the platform or the pedestal
holding the said object. Secondly, by comparing peak of the
object with the peaks in the histogram associated with at
least two reference objects in the camera image as shown in
step 525. At step 530, the position of the peak corresponding
to the object is mapped to a moisture value. This could be
done in different ways. First, mapping the position of the
peak to the moisture value based on a function as shown in
step 535. Second, by interpolating the moisture value of said
object using a database using a plurality of data points;
relating the position of peaks with moisture values as shown
in step 540. Third, by performing said mapping by a neural
network trained based on the pre-calibration as shown in step
545.

Summarizing, the present invention introduces an apparatus to
determine the moisture content of an object. This is done
using a histogram computed from an infrared camera image of
the object. The histogram shows a peak associated with the
object, whose position in said histogram is first identified.
The apparatus further maps the position of the peak to a
moisture value. The use of infrared image and histogram
information for determining the moisture makes the whole
process a fast and non-destructive process for moisture
detection. .
The resulting advantages of the idea are many-fold: 1) A
camera has a field of view; hence can measure the moisture
content of the sample object as a whole, as opposed to that
of a specific location on the sample, leading to a more
holistic picture. 2) This is non-invasive technique. 3} It is
a non-destructive technique. 4) Results as to the indication
of the moisture value could be obtained instantaneously. 5}
Operator training is easy. 6} Results are not influenced by
external environmental conditions.
Although the invention has been described with.reference to
specific embodiments, this description is not meant to be
construed in a limiting sense. Various modifications of the
disclosed embodiments, as well as alternate embodiments of
the invention, will become apparent to persons skilled in the
art upon reference to the description of the invention. It is
therefore contemplated that such modifications can be made
without departing from the spirit or scope of the present
invention as defined.

Patent claims:
1. An apparatus (100) for determining moisture content of an
object (110), comprising:
- infrared camera (120) for capturing a camera image of the
object (110);
- computing means (130) to compute a histogram (200, 400) for
the camera image;
- identifying means (140) to identify a position of a peak
(230, 430) in the histogram (200, 400) corresponding to the
object (110); and
- means (150) for mapping the position of the peak (230, 430)
corresponding to the object (110) to a moisture value.

2. The apparatus according to claim 1, wherein identifying
means (140), identifies the position of the peak (230, 430)
in the histogram (200, 400) based on a known relation between
the temperature of the object (110) and surrounding of said
object (110) .
3. The apparatus according to claim 1 or 2, wherein
identifying means (140) identifies the position of the peak
(230, 430) in the histogram (200, 400) by comparing peaks in
the histogram associated with at least two reference objects
(320, 330) in the camera image.
4. The apparatus according to any of the preceding claims,
wherein the at least two reference objects (320, 330) in the
camera image has different known temperatures.
5. The apparatus according to any of the preceding claims,
wherein the means (150) for mapping, maps the position of the
peak (230, 430) to the moisture value based on a function.
6. The apparatus according to any of the preceding claims,
wherein the function is derived based on a pre-calibration
done for a specific type of object.

7. The apparatus according to any of the preceding claims,
wherein the means (150) for mapping further comprise a
database with a plurality of data points relating the
position of peaks with moisture values from which the
moisture value of said object (110) is interpolated.
8- The apparatus according to any of the preceding claims,
wherein the means (150) for mapping comprises a neural
network, said neural network trained based on the pre-
calibration.
9. A method of determining moisture content of an object
(110), comprising the steps of:
- capturing a camera image of the object (110) using an
infrared camera (120);
- computing a histogram (200, 400) for the camera image;
- identifying a position of a peak (230, 430) in the
histogram (200, 400) corresponding to the object (110); and
- mapping the position of the peak (230, 430) corresponding
to the object (110) to a moisture value.

10. The method according to claim 9, wherein identifying the
position of a peak (230, 430) in the -histogram (200, 400) is
based on a known relation between the temperature of the
object (110) and surrounding of said object (110).
11. The method according to claim 9 or 10, wherein
identifying a peak (230, 430) in the histogram (200, 400)
further comprise the step of comparing peak of the object
(110) with the peaks in the histogram associated with at
least two reference objects (320, 330) in the camera image.
12. The method according to any of the claim 9 to 11, wherein
the at least two reference objects (320, 330) in the camera
image has different known temperatures.

13. The method according to any of the claim 9 to 12, further
comprise the step of mapping the position of the peak (230,
430) to the moisture value, based on a function.
14. The method according to any of the claim 9 to 13, wherein
the function is derived based on a pre-calibration done for a
specific type of object.
15. The method according .to any of the claim 9 to 14, wherein
mapping the position of the peak (230, 430). further comprise
the step of interpolating.the moisture value of said object
using a database with a plurality of data points relating
the position of peaks with moisture values.
16. The method according to any of the claim 9 to 15, wherein
mapping the position of the peak (230, 430) corresponding to
the object (110) to a moisture value involves performing said
mapping by a neural network trained based on the pre-
calibration.

The present invention introduces apparatus (100) to determine
the moisture content of an object (110). The apparatus (100)
comprise an infrared camera (120) directed towards the object
(110) to capture a camera image. The apparatus further has a
computing means (130) to compute a histogram (200, 400) for
the camera image and an identifying means.(140) to identify a
position of a peak (230, 430) in the histogram (200, 400)
corresponding to the object (110). The apparatus has a means
(150) for mapping the position of the peak (230, 430)
corresponding to the object (110) to a moisture value.