Está en la página 1de 84

19 March 2014

Image Quality is
Worthless
without Usability.
A lesson in how to make video surveillance more usable for the
end user

www.axis.com
Objectives

> Become familiar with


– Term “image usability”

> Understand
– Key factors required to obtain the most
useful images
– Impact of light

> Be able to
– Modify camera settings to compensate for
various lighting conditions

www.axis.com
What is Image Usability?
There is nothing wrong with the quality of this image

www.axis.com
“Eighty percent of
surveillance video
submitted to the police
is unusable!”
www.axis.com
Source: Swedish Police Central Imaging Group White Paper on Image Usability
Image usability starts with....

1 – Define the purpose


2 – Determine the resolution
3 – Determine how cameras use and are affected by light
4 – Determine camera placement
5 – Select and configure the camera

www.axis.com
Define the purpose of the surveillance

> Why surveillance?


> What level of detail is required?
> Where do you need the required level of
detail?
> Which area needs to be covered?

www.axis.com
Goal definition – Deterring or discreet?

Is deterrence a factor... ...or is discreet surveillance a priority?

www.axis.com
What is your surveillance need?

Identification?
Detection?
Recognition?

www.axis.com
Goal definition – Overview or high detail?

Is overview enough... ...or is high detail required?

www.axis.com
Aspect ratio – Relationship between horizontal and vertical

4:3 :16
16: 9
www.axis.com
Axis’ Corridor Format 9:16

> Usable in aisles and corridors where lots


of image information is otherwise
redundant
> A way to increase pixel density in tall,
narrow scenes
> Need visual evidence?

www.axis.com
Exercise – 3.1 Axis’ Corridor Format

> In this exercise you will set your camera to


corridor format and see how this affects the
field of view
> Upon completion you should be familiar with
– Benefits of using Axis’ Corridor Format

www.axis.com
Image usability starts with....

1 – Define the purpose


2 – Determine the resolution
3 – Determine how cameras use and are affected by light
4 – Determine camera placement
5 – Select and configure the camera

www.axis.com
What is a pixel?

Pixel

www.axis.com
Let’s enhance

www.axis.com
Let’s Enhance!
Resolution is determined
when video is captured –
so make sure to capture
your video in an
appropriate resolution!
www.axis.com
Megapixel resolution & HDTV capabilities

Resolution comparisons Aspect ratios

HDTV 1280 x 720

HDTV 1920 x 1080

5 Megapixel 2560 x 1920

www.axis.com
Resolution comparisons in the real world

1 MP 3 MP 5 MP
www.axis.com
Megapixel resolution & HDTV capability comparisons
With HDTV 1080p and 20x zoom,... With HDTV 720p and 18x zoom,...

...the license plate of a car ...the license plate of car


275 m (900 ft.) away can be read 200 m (660 ft.) away can be read
www.axis.com
Megapixel resolution & HDTV capability comparisons
With D1 and 18x zoom,... With 4CIF and 12x zoom,...

...the license plate of a car ...the license plate of car


140 m (460 ft.) away can be read 76 m (250 ft.) away can be read
www.axis.com
Pixel density and scene width

Level of detail Pixels/face Pixels/m


Identification
80 500
(challenging conditions)
16 cm
Identification
40 250
(good conditions)
Recognition 20 125
Detection 4 25

www.axis.com
Pixel density and scene width
How wide a scene can a 1280x720 camera cover?

Level of detail Scene width


Identification
1280/500 = 2.56 m
(challenging conditions)
Identification
1280/250 = 5.12 m
(good conditions)
Recognition 1280/125 = 10.24 m
Detection 1280/25 = 51.2 m

www.axis.com
Let’s do the math!

80 cm
16 pix

100 cm @ 80 pixels
> A “average” face is 16 cm wide
> Recommendations for positive identification of a face require 80 pixels in challenging conditions
> 80 pixels / 16 cm = 5 pixels / centimeter
> 5 pixels/centimeter x 100 cm = 500 pixels / meter

www.axis.com
Delivering usable images at a specific point – Capture line

Where does the


customer need the
capability?
www.axis.com
Customer requirement:

“As a high school


principal, I want to be
able to identify students
roaming the hallways
during class!”

www.axis.com
Customer requirement example - metric
Lens 8 mm Lens 3 mm
“As a principal, I’d want to be able to identify
Capture line for 80 pixels Capture line for 80 pixels
• Width 2.56 m
• Width 2.56 m
students roaming the hallways during class!”
• Distance 5.65 m • Distance 2.12 m
End of hall - 45 pixels End of hall - 17 pixels

www.axis.com
Exercise – 3.2 pixel calculation

> In this exercise you will calculate the possible


scene width for
detection/recognition/identification, using
different resolutions
> Upon completion you should be familiar with
– How resolution and pixel requirements affect
the scene width

www.axis.com
DEMO – Pixel counter

> The pixel counter is a valuable tool for


ensuring a certain surveillance level at a
specific capture line
> This demo will show an example of how the
pixel counter can be used

www.axis.com
Image usability starts with....

1 – Define the purpose


2 – Determine the resolution
3 – Determine how cameras use and are affected by light
4 – Determine camera placement
5 – Select and configure the camera

www.axis.com
How does the camera work?

Think photography!

From Greek:
• Photos (light)
• Graphos (drawing)

www.axis.com
Too much can be as bad as not enough!

Overexposed

Underexposed

www.axis.com
Aperture and shutter – the interdependent image controls

Aperture

Shutter

www.axis.com
Aperture and shutter – Interdependent image controls

Aperture

Shutter

www.axis.com
Fast shutter example
> Fast shutter – short exposure per frame
> Frame rate kept at full speed
> No noticeable image defects
> Movements are crisp with solid edges
> Fairly shallow depth of field

www.axis.com
Slow shutter example

> Slow shutter – longer exposure per frame


> Frame rate reduced
> Blurriness around moving objects
> Extended depth of field

> Choppy video because of lower frame rate

www.axis.com
Fast shutter – Low frame rate

> Fast shutter – short exposure per frame


> Frame rate dropped to 6 per second
> Frame rate, not shutter speed the cause of
choppy video
> Movements are crisp with solid edges

www.axis.com
Exercise 3.5 – Motion blur

> In this exercise you will learn how shutter


speed affects image usability
> Upon completion you should be familiar with
– Visual effects of different shutter speeds
– Shutter speed settings
– How image quality priorities work in Axis
cameras

www.axis.com
Fluorescent lights can cause poor image usability

www.axis.com
How to set flicker-free settings

> Initial configuration requires


the power line frequency be
set to the region
> This does not affect the
shutter capability in the
camera
> It is still possible to have
image flicker, even though
this setting is correct

www.axis.com
How to set flicker-free settings

> Image flicker can be further


compensated by the camera
exposure settings
> The setting by default is
“Automatic”
> When this setting is changed,
notice what happens to
shutter speeds:
– 60 Hz mains =1/120th
second
– 50 Hz mains = 1/100th
second

www.axis.com
Problem solved

www.axis.com
Visible light

www.axis.com
Color temperature
Kelvin color temperatures
> Color temperature is measured in degrees Kelvin (K) 10,000 K
Sky –
and visible light ranges from 1,500 K to 10,000 K Partly cloudy
9,000 K
> Daylight color temperature shifts during the day
8,000 K Outdoor shade area
> Indoor light color temperature depends on source
7,000 K Overcast
Daylight
> Human eyes AXIS White Light LED

6,000 K
– Will compensate for the differences in color
Daylight fluorescent tubes

> Cameras 5,000 K Noon daylight

– Must be adjusted for illumination through white


Summer daylight
4,000 K
balance
Warm fluorescent tubes
3,000 K
– Usually automatic modes that calculate the white 40W incandescent bulb

balance automatically 2,000 K High-pressure sodium lamps

Candle

1,000 K
www.axis.com
White balance

> To achieve the right color representation


> Neutral (black/gray/white) colors in the scene should stay neutral in the
resulting image, regardless of the illumination

www.axis.com
White balance

> Selecting white balance options


allows for adjusting the image
to better represent the
environment
> Under the white balance drop
down menu are various options
each representing a different
color temperature

www.axis.com
White balance – Manual

> Selecting “Manual” white


balance allows for the
placement of an index
point on a specific area of
color interest to match in
the image
> ‘Edit’ allows for the
sizing and placement of
the configuration window

www.axis.com
White balance – Custom

> Selecting “Custom” white balance


allows for the placement of a
window over a specific area of
color interest in the image
> “Edit” allows for the
sizing and placement
of the configuration
window

www.axis.com
Light is the biggest challenge for usable surveillance
> Light direction - Color temperature - Light intensity - Dynamics

www.axis.com
Exercise 3.7 – White balance

> In this exercise you will learn to set white


balance
> Upon completion you should be familiar with
– How to set camera white balance according
to available light sources

www.axis.com
Light intensity

Direct sunlight: 130 k lux Office: 200 lux Shopping mall: 500 lux
Shadow 10 k lux Indoor floor 50 lux

www.axis.com
How much light does a surface reflect?

Surface Reflection, %
Dark asphalt 5
Dark clothes 10
Grass 15
Gray colors 50
Yellow colors 60
Aluminum 70
Snow 80

www.axis.com
Adding external light

Advantages Disadvantages
Visible deterrent Light pollution
Quick start (LED) Reduced distances
White light
Full color rendition
Easy setup
Covert or Semi-covert Limited deterrent
Infrared Longer distances More difficult to set up
No light pollution

www.axis.com
Camera location – Adding light

> Match the angle of light to the angle of


view provided by the camera
– Too narrow will leave dark areas at each
side
– Too wide will leave areas in the distance
too dark
> Ensure a distance of at least 1 m (3 ft)
between camera and light source to avoid
problems with insects
> For light with adjustable angle of coverage:
Wide angle distance approximately 66 % of
max distance

www.axis.com
The inverse square law of light

Double the distance –


four times the light!

Light
source

1m 2m 4m 8m
64 lux 16 lux 4 lux 1 lux
1 4 16 64
www.axis.com
Ambient low light outdoors
Color mode Night mode

www.axis.com
Auxiliary infrared light
IR LED illuminators Overexposed IR LED illuminators

www.axis.com
Lightfinder technology

> Axis’ Lightfinder technology is a


combination of many factors:
– Axis’ in-house chip technology
– Latest-generation, light-sensitive sensor
– Optimized, high-quality lens
– Axis’ know-how and powerful image
processing capabilities

www.axis.com
Image example – Extraordinary low-light sensitivity
AXIS 221 AXIS Q1602

0.5 lux, 30 fps


www.axis.com
WDR – What is Wide Dynamic Range?

> Ratio of the brightest point of an


image to the darkest point of the
same image
> Goal is to get details in the whole
image
– Specifically for strong backlight
scenarios

> Measure is relative

www.axis.com
Why WDR?

> Provide detailed images of


complicated scenes with extreme
illumination
> Improve visibility of details in
blinding situations
> Improve color and overall image
fidelity
> Fully exploit the dynamic details in
a scene

www.axis.com
AXIS Q1604 – WDR dynamic capture

AXIS Q1604 viewing open doorway

AXIS Q1604 viewing a halogen work light

www.axis.com
The challenge – Varying light levels

This image is underexposed, when This image is overexposed, when


trying to capture the environment trying to capture the environment
outside indoors
www.axis.com
The challenge – Varying light levels

Neither image manages to capture all


the details both outside and indoors
www.axis.com
What we want...

An image that shows the details


in all areas – at the same time
www.axis.com
WDR results – Tunnel example

www.axis.com
AXIS Q1604 in headlights

www.axis.com
WDR artifacts in general – Ghosting and motion blur

> Ghosting

> Motion blur

www.axis.com
Camera settings – WDR mode

> Capture modes 50 Hz and 60 Hz


with 25/30 fps
– To reduce flickering with WDR

> Less customization in exposure


priority when WDR mode is on
> WDR – dynamic contrast in
PlainConfig

www.axis.com
Exercise 3.8 – Wide Dynamic Range: Dynamic contrast

> In this exercise you will learn to set Wide


Dynamic Range: Dynamic contrast
> Upon completion you should be familiar with
– Benefits of WDR

www.axis.com
No light – Thermal camera
24/7 detection in challenging conditions

> Detection in the dark


– No energy-consuming floodlights

> Detection in difficult daylight conditions


– Bright light
– Shadows
> Other challenging conditions:
– Rain
– Snow
– Haze
– Smoke

www.axis.com
No light and light? Combine both!

Dual camera on single PT control


> Perimeter protection with thermal and
conventional camera

Dual camera on single PT control


 Perimiter protection with thermal and convential camera
 Ideal for government

Detect with AXIS Q1921 Identify with AXIS Q1755

www.axis.com
Image usability starts with....

1 – Define the purpose


2 – Determine the resolution
3 – Determine how cameras use and are affected by light
4 – Determine camera placement
5 – Select and configure the camera

www.axis.com
Placement greatly affects image usability

> Out of reach for vandals


> Legal aspects
> Light properties
> Field of view
– Line of sight
– Height and vertical angle
– Coverage

www.axis.com
Camera mounting – Mounting height
High up... ...or face level?

Consider the risk of tampering or vandalism!


www.axis.com
Camera location – Unsuitable angle

www.axis.com
Camera location – On the sunny side

> Consider the direction of the sun when installing cameras

www.axis.com
Image usability starts with....

1 – Define the purpose


2 – Determine the resolution
3 – Determine how cameras use and are affected by light
4 – Determine camera placement
5 – Select and configure the camera

www.axis.com
Evaluate trade-offs – Which is “better”?

> Consider the trade-offs


in your solutions
> Prioritize
> Realize that the “best”
solutions cost money

www.axis.com
Understand: Wide angle and 360°/180° panoramic views
Camera overview
Quad views

Images taken with AXIS M3007-PV


www.axis.com
Understand: Wide angle and 360°/180° panoramic views

Single panoramic Double panoramic

www.axis.com
Demo: Axis product selector and lens calculator

www.axis.com
Conclusions

> Chose pixel density, preferred format and


aspect ratio to fulfill the surveillance need
> Light properties to consider
– Light direction
– Color temperature
– Light intensity
– Dynamics

www.axis.com
www.axis.com

También podría gustarte