Image processing technology is widely used in the manufacturing and logistics industries, and has become indispensable in modern industries where automation is essential. In addition, the field of application is expanding further, and it is beginning to be used in a wide range of fields such as medicine and agriculture. We will introduce what image processing technology is and what benefits it brings, as well as an overview and examples of its use.
Overview and types of image processing technology
Image processing technology can be summed up in a nutshell, but there are many types of processing, and the results obtained vary depending on the type of processing. Let's look at the types of image processing while giving an overview of what image processing is.
What is image processing?
Image processing is a technology that classifies image data into patterns that meet certain conditions, compares them with reference data, and processes them. Convert, transform, analyze, and extract information from the acquired image.
In addition, the technology from image acquisition to judgment and discrimination of data obtained by image processing is called machine vision.
In this way, the scope of work increases in the order of image acquisition, image processing, and machine vision.
Reasons why image processing technology has spread
It goes without saying that product quality is important in the manufacturing industry. Visual inspection is the most common method for judging the quality of products, and is performed at many quality inspection sites.
While FA (factory automation) has been promoted in most manufacturing processes, it has been considered difficult to automate the visual inspection process. This is because it was difficult for machines to take over the work of judging products by looking at them with human eyes, especially in the "judgment" part.
However, the major change is the practical application and spread of AI. With the establishment of a machine learning method called deep learning, AI has greatly improved its accuracy and made great strides toward practical application. It can be said that AI has become able to make judgments that only humans could do, and the environment is now ready for the introduction of inspection automation in various fields.
The realization of such AI-based automation has made it possible for computers to perform image processing instead of humans. It has been introduced from the manufacturing industry, where automation is strongly required, and has been applied to a wide range of fields such as logistics, agriculture, and medicine.
Image Processing type
Image processing does the following:
- Matching
Shape-based matching, feature point matching, etc.
- High Precision Measurement
High-precision measurement in fine pixel units
- Color Processing
Color classification, matching using color information, various color system conversions, etc.
- Deep Learning
Inspections that could not be handled by the rule base
- Alignment
Fast, stable and accurate alignment even with low contrast
- Character Recognition
Automatically detect and read character information from images even from a few lines
- Visual inspection
Responding to difficult inspection items
- Calibration
Removal of lens distortion, conversion to the world coordinate system, connection of multiple cameras, etc.
- Barcode/Datacode
Reading various codes, automatic learning estimation of parameters, etc.
- Non-defective product inspection
Detect the difference from non-defective products
- 3D processing
Surface matching, 3D data processing, 3D display, etc.
- Filter
Autofocus, high dynamic range, optical flow, etc.
- Taxonomy
Subtle classification close to human sensibility by software
- Blob analysis and morphology processing
Ultra-fast region manipulation, labeling, morphological processing, etc.
- High Speed Processing
Automatic parallel computing, multithreading, programming, etc.
Advantages of introducing image processing technology
The introduction of image processing technology has the following advantages.
Automation by realization of labor saving and manpower saving
Work is labor-saving by utilizing image processing technology. By accumulating labor saving, it leads to labor saving and automation can be achieved.
Automating it will lead to elimination of labor shortages and reduction of personnel costs.
Work standardization and independence
By using machines, there is no variation in standards, and processing and judgment are based on fixed standards. Human error does not occur.
In addition, standardization is realized by unifying standards. Since the same result can be derived regardless of who operates it, it is also possible to achieve non-personalization.
Improving inspection accuracy and speed
Continuous processing by machine becomes possible, and high-speed inspection becomes possible. Accurate work by machines also improves accuracy.
Improving the working environment and improving business continuity
It is possible for machines to take over work in dangerous places and in environments that are harmful to the human body.
It also eliminates long working hours and improves work-life balance. By using machines that do not require holidays as a labor force, it becomes easier to respond to sudden vacancies, and it is possible to build a flexible work system.
In this way, the ability to flexibly assign personnel and build a labor system will lead to an improvement in the overall working environment.
Realization of DX
DX will be realized by having image processing systems take over what people used to do.
This is more than just an efficiency effect. The construction of new work processes using digital technology creates the potential for innovation and the creation of new business models.
Application Examples of Image Processing Technology
Image processing technology is used in the following situations.
Labeling products on conveyor
Image processing technology is applied to the task of attaching labels to packaged frozen foods.
As frozen foods flow on the conveyor, the individual angles at which they are placed make a slight difference. Therefore, in the past, labels were affixed to packages manually by humans. However, by introducing a robot capable of image processing, we succeeded in automating this work.
Not only did we succeed in automating the process, but we were also able to speed up the work and improve the pasting accuracy.
Detecting molded products in molds
In injection molding of plastics, the product is molded while the mold is closed and ejected when the mold is opened. There are several ways to take out this product, such as robot take-out and automatic drop ejection.
At this time, if the next molding cycle starts with the product not being removed from the mold, the machine will stop and it will take a long time to recover. In addition, it can also damage the mold.
In order to prevent such troubles, we use an image sensor to detect the inside of the mold to check if there is any remaining molded product, realizing safe and stable production.
Optimal cutting of wood
When cutting wood from logs, the challenge is how to cut it in a way that is efficient and yields a high yield.
It is possible to optimize the extraction by performing cross-sectional and volumetric measurements by photographing from all around and acquiring highly accurate shape data.
Car number recognition system
It is possible to recognize the license plate of a car, detect the size of the plate, apply 3D correction, and convert it into an image that looks like it is viewed from the front. From here, the car number can be read by character recognition and automatically registered in the database.
It is used for entry/exit management in hotel parking lots and automobile monitoring systems on highways.
Inspect food packaging labeling
Information such as the product name, raw materials, processor, manufacturer, date of manufacture, etc. is displayed on food packaging. In addition, prices and barcodes may be entered.
If the information in these letters and codes is not displayed correctly, the necessary information cannot be conveyed to consumers. It also leads to the possibility of errors in machine reading during logistics and sales.
It is possible to check whether there are any errors in these prints by performing character recognition and comparison with reference patterns.
Points to Note When Introducing Image Processing Technology
When introducing image processing technology, the following points are important.
Choose a trusted integrator or vendor
There may be many cases where you are fumbling about what kind of preparations you should proceed with when introducing an image processing system. There is a possibility that demerits will increase, such as reaching irrelevant results even if you proceed with the introduction with only in-house personnel while still fumbling, or purchasing unnecessary equipment and increasing costs.
To minimize this kind of waste, it is best to consult an expert who specializes in image processing technology.
Select an integrator or vendor who understands your company's issues and requirements and can make optimal proposals.
Select a camera suitable for the site environment
It is important for image processing equipment to be durable for industrial use and to be able to supply parts for a long time.
In addition, since embedded systems can realize specific functions, they have many advantages and are easy to introduce. In this way, choosing a camera that is durable enough to handle the environment in which it will be used and that specializes in image processing functions is advantageous in terms of cost and performance.
About the embedded system, please refer to the commentary on what is an embedded system and its advantages and points to consider when introducing it.
Select image processing software suitable for your environment
When introducing a new system, select one that can use the data that your company has accumulated so far. By choosing a system that can work with existing data, you can use existing databases and customer information.
This leads to minimization of time and man-hour costs in installation work.
Standardization of operation procedures and preparation of manuals
We will standardize the operation procedures and prepare them as manuals, and promote the education of employees. At the same time, we will decide how far we will respond in-house and where we will contact the vendor when a problem occurs.
The contact method should also be documented.
Image processing technology will expand to a wider range of fields in the future
Regarding image processing, we introduced what kind of technology can be used, in what situations it is used, and what is important when introducing it.
Image processing technology has great potential due to the
practical application of AI. Machines can now make judgments and judgments that
could only be done by humans, making it possible to effectively use the data
obtained by image processing technology. As a result, image processing
technology, which was mainly used in the manufacturing and logistics
industries, is now being used in a wide range of fields. It is also applied in
the agricultural and medical fields, and is an important elemental technology
for next-generation technologies such as autonomous driving.