About technology used in vision applications
Now, vision applications are often seen as something very typical and present in almost every factory in the world. Today, I would like to introduce you to several unusual vision applications that can change your view on vision application market and broaden your view on the technology that surrounds us.
With the development of technologies, such as multispectral vision, deep learning, thermovision, high-speed cameras and many more, scientists were able to develop revolutionary solutions. The industries, in which vision systems are now used, can often be surprising:
- Space industry
- Traffic organisation
- Public safety
- Environmental protection.
Unusual vision application examples
Space industry — processing of hyperspectral images in orbit
Let’s start with something out of this world. Intuition-1 will be a 6U-class satellite with data processing unit, enabling on-board data processing, acquired via hyperspectral instrument with spectral resolution in the range of visible and near infrared light. The hyperspectral camera will use up to 150 spectral bands in the range of 470 nm – 900 nm. Every scene captured by the camera will contain an image divided into multiple frames. Each frame will register a different spectrum range.
Intuition-1 satelite from KP LABS.
Thanks to deep learning algorithms, Intuition-1 can have a wide range of uses. For example, it is used in agriculture for:
- Land coverage classification
- Crop forecasting
- Crop maps
- Soil maps
- Plant disease detection
- Biomass monitoring
- Weed mapping.
Image from satelite in agriculture applications.
In-orbit data processing can be also used in sectors, such as forestry and environmental protection. Examples of using the satellite in these sectors are:
- Forest classification
- Identifying species and the condition of forests
- Forestation planning
- Water and soil pollution maps
- Land development management and analysis.
As you can see, deep learning, combined with multispectral vision and space engineering, can be used to make many amazing applications. You can also read more about this solution under this link.
Sport — machine vision c
These days, machine vision cameras could be used everywhere — even in sport and entertainment. For example, they are used to support or even replace brodcast cameras, track objects like ball or players, helps in stadium and crowd management. Data acquired and measured with machine vision cameras coud be also used for overall game analysis, including player performance analysis. In this article I will focus on the 3D replay systems.
JAI’s machine vision camera used in the 3D reply systems. Image source.
Using JAI’s high speed cameras allowed to create amazing stereoscopic 3D views of the action. A typical live broadcast has been enriched with new 3D replay system, which allowed to obtain surprisingly good effects. In the source article, several transmission support systems have been described. The most advanced one was built based on 42 cameras mounted around the stadium. Advanced image processing allowed to reconstruct stereoscopic 3D views.
Public safety — vision body temperature measurements
A while ago I wrote about methods of coping with coronavirus pandemic and I was presenting mask detection. In addition, today, I’m going to write somthing more about thermal cameras and systems, that allow to measure body’s temperature. Since I wrote about Dahua’s system, many solutions, based on existing thermal cameras and thermal sensors, appeared on the market. Moreover, several new devices, specially designed for temperature measurement, have also been released, for example — A400/A700 from FLIR.
FLIR’s Thermal Smart Sensor Camera. Image source.
Those systems become more and more popular, however, a few months ago such solutions were considered unusual. There are many solutions on the market, but the most accurate of them use AI to detect the area around the eyes. In combination with devices providing constant thermal reference point (so-called black body), they allow you to achieve accuracy of ± 0.3 ℃. In my opinion, never before have people noticed the great advantage that vision systems can give, compared to traditional inspections and measurements.
Let’s start from two interesting things, like fuit detection and robotic mobile platform. It becomes even more impressive when you combine it into really unusual vision application — robotic harvesting. Solution from Metomotion company uses not only 3D imaging, but also really advanced machine vision algorithms. GRoW (Greenhouse Robotic Worker) in not only used for harvesting — it is also a multipurpose robotic system for labor-intensive tasks in greenhouses. It can be used for pruning, monitoring, pollinating as well.
At this point, I would like to show you ready to use deep learning tool for the point location. In the video above you have seen how to use it for bees localization. The goal of the complex application was to check, whether bees are infected by varroosis or not. The parasite attaches to their bodies and causes a characteristic red inflammation, which can be detected. As a result, bees are classified according to their health condition. Adaptive vision’s tool can be used for many other machine vision and agriculture unusual vision applications. For example, in automotive, food production, for locating and counting fruits and vegetables, and for applications like this described in the next paragraph.
Environmental protection — bird control solutions
Automated bird repellent system from Bird Control Group. Image source.
Summary and conclusions