Biz-AI Solution
Among the fields that increase productivity and improve competitiveness by applying smart factory/smart farm equipment and technology
to the production sites of small and medium-sized manufacturing companies,
they are currently conducting business in three fields related to AI solutions.
-
Smart Factory/Smart Farm AI Solution
It is an automatic inspection system that can detect defects in real time using a line scan camera and AI machine vision software.
It collects process data from the manufacturing site, analyzes and visualizes it, monitors the process status in real time, and analyzes the collected data in bulk.
It is a system that can find the cause of defects by analyzing big data.. -
Facility predictive maintenance AI solution
As a facility management solution that detects abnormal conditions of facilities and predicts how they will lead to failures with AI, and presents solutions, key data (facility temperature, vibration, current, etc.)
It is a system that predicts the actual occurrence of an error by finding a correlation between error occurrence and error occurrence. -
Animal Behavior Observation AI Solution
It is a solution that recognizes the current state by observing the animal's behavior through video and analyzes it with AI to predict abnormal conditions or abnormal behavior.
It is a system that enables the farmer to respond by identifying various signs such as the estrous behavior of dairy cows or Korean cattle in the barn.
1. Smart Factory AI Solution
Real-time high-speed AI defect detection system using line scan camera
It is an automatic inspection system that can detect defects in real time using a line scan camera and machine vision software.
It is a system that detects defects in real time from continuously moving objects.
Including PCB and semiconductor parts, it is possible to detect defects in the manufacturing sector of materials
such as plastic, paper, foil, film, and metal.
AI machine vision systems replace the process of human eyes and judgment in quality inspection
and inspection work during the manufacturing process, enabling precise and fast processing even in harsh environments.
It utilizes the line scan method, not the area scan method that analyzes an object by photographing it in a still state.
Moving objects can be quickly photographed in real time, and large objects and high-resolution images can be processed quickly.
In order to recognize irregularly supplied products and provide accurate shooting timing to the camera, we developed a technology that can give a signal (trigger) for camera shooting. A webcam was installed at the part where the subject entered, and when the front part of the product was recognized through the webcam, the current moving speed was measured, and a trigger was generated reflecting the distance to the camera to give an appropriate shooting timing.
For more sophisticated defect detection, defect detection software using AI deep learning was developed, and pattern inspection of the existing image processing method is also applied to conduct efficient vision inspection.
Apply an unsupervised learning method that learns only standardized non-defective images and considers non-defective products as defective and finds them, or applies a supervised learning method that labels and learns all types of defects separately and detects defects by type.
2. Facility predictive maintenance AI solution
Facility predictive maintenance AI solution detects abnormal conditions of facilities,
diagnoses and identifies whether it is caused by a defect in the facility,
is a temporary phenomenon, or is a problem in the facility itself,
and then predicts how it will develop and lead to failure. It is a skill.
Data sensing processing detects key state values (temperature, humidity, PH concentration, vibration, current, etc.) of process equipment or working materials with precision sensors and transmits them to the internal receiver.
And the transmission kit transmits the received data to the external main server.
At this time, the power supply and network are processed wirelessly as much as possible to minimize interference with the work.
The collected sensor data, facility management history data, and facility basic specification data are all collected and analyzed in real time, and the analysis results for the current facility status are visualized.
In addition, real-time monitoring is supported by utilizing the cloud web, and when the set threshold is exceeded, the manager is notified immediately so that he can respond.
It is a system that analyzes the status data (temperature, vibration, current, etc.) of major facilities in real time to detect anomalies in advance and predict failures.
Existing well-known equipment has data secured in case of failure, but in the case of new equipment, there is no data on abnormal operation.
Therefore, it is necessary to learn and predict all of the numerous patterns for normal motions, and at some point in time, if a movement in a state that is not the normal pattern predicted by the system is detected, it includes a function to determine it as an abnormal motion and notify it.
3. Animal Behavior Analysis AI Solution
Accurately detecting the estrous state is the most important factor for planned mating of Korean cattle,
and the fertility rate of breeding cows can be improved by detecting the estrus state of the cow.
Therefore, it is common in most farms to detect estrus by detecting estrous behavior of ovaries in estrus through visual inspection
of CCTV recorded screens, but this behavior may occur at any time of the day, usually from 12:00 p.m. to 6:00 a.m.
There is a concentration of estrus between poems. In this situation,
it is very difficult for a person to visually observe estrous behavior 24 hours a day,
and it is more difficult to visually observe estrous behavior that occurs at night than during the day.
In order to solve this problem, we developed a 'image big data (deep learning) based estrous behavior detection algorithm'
using CCTV images already installed in livestock farms.
Features
-
Superior
Application of superior
unique AI technologyCan apply AI technologies
in multiple areas. -
Flexible
Flexible adaptability
to working environmentCan make automation even in
non-standard working environment. -
Easy
Easy
developmentCan be developed easily
by the users. -
Fast
Fast outcome
supportsProvides working and quality
information real-time basis.