Anhui Keye Intelligent Technology Co., Ltd

IML Cup AI Vision Inspection Machine proceed by image acquisition, image processing, image annotation, Al algorithm models, and software scheduling to inspect apparent defects of IML products in industries such as food, dairy, beverage and more industries. lt rejects defective products from line instantly and counts good products for box division.    Product Parameters        What product can be inspected? IML cup/Label print - in-mold labeling for food, dairy products, beverages,etc         Defects Detection         Inspection Theory    Inspection process of Keye Tech vision inspection system to detect appearance defects: Feeding device: Products transfer to a workstation via feeding devices, entering the inspection machine orderly one by one. Image acquisition: High-pixel images are captured by customized system of self-developed industrial cameras and LED light sources. Image processing: Captured images upload to our self-developed KVIS COULD AI Platform for few image annotation and model training(within 1h). AI algorithm models: AI algorithms analyze the processed images, sort products into "NG" (defective) and "OK" (qualified), and trigger inline rejection of defective products via pneumatic devices. Software scheduling: The software provides sample comparison standards, allowing customers to select the most suitable criteria to balance accuracy and defect rates. Counting and box division: Defective products are instantly rejected into a separate box. Good products are counted for box division.    10 Cameras Standard Solution(customizable)    Station1 1 Camera to check cup bottom defect Station2 1 Camera to check cup top flat defect Station3 4 Cameras to check cup outer sidewall defect Station4 4 Cameras to check cup inner sidewall defect QAs of IML Product AI VISION INSPECTION SYSTEM Q: What’s kind of product can be inspected? A: It can inspect IML products, IML cups and containers. Q: What kind of defects it can detect? A: All the defects that can be seen by eyes is able to be inspected, but it is only for appearance defect inspection. Q: How many cameras are equipped in the machine? A: The camera quantity is also decided by the detected defect kinds and positions, standard configuration is 10pcs. Q: What’s the definition of the camera? A: The definition of the camera and lens are depended on client’s precision request, i.e, 0.1mm*0.2mm, 0.5mm*0.5mm. Q: If the camera is self-developed or buy from the professional camera producers? A: Most of the cameras are designed and developed by ourselves, the lens and sensors are supplied by the worldwide brands. Q: Is the cameras are the most important for the inspection? A: Camera is one of the most important hardware of the IML inspection machine, the AI algorithm software system is the key technology for the machine. Q: How to solve installation and adjustment? A: We will send our skilled engineers to you for installation and adjustment, and customer can also send their engineers for training. Q: Is difficult to use your machine, how to train our people? A: The system is sophisticated while the operation is easy due to our high level AI system. We can train your people during the installation or by remote control, you can also send your technician to our company to learn, it’s free.

PET Preform AI Visual Inspection System conducts visual defect detection on water, beverage, oil and other preforms via a systematic workflow: image capture, intelligent image processing, automated annotation, advanced AI algorithm modeling, and software coordination. It instantly rejects non-conforming products, automatically counts qualified units, and supports batch box division for streamlined production.              What product can be inspected? PET Preform - water preform, beverage preform, oil preform, etc.    Defects Detection         11 Cameras Solution (customizable)      7 cameras solution: 1 for top mouth, 1 for bottom, 1 for supporting ring, 4 for body sidewall. 11 cameras solution: 1 for top mouth, 1 for bottom, 1 for supporting ring, 1 for injection gate, 4 for neck thread, 4 for body sidewall. More customized solution available.    Required Information Before Quotation    1.Integration method: offline/in-line inspection 口 Offline inspection: The inspection is separated from production line (preform feeder + inspection machine). Please advise production capacity. 口 In-line inspection: the preforms feeder connects to your production line, then enter inspection machine. Please advise line speed. 2. Preforms Size Info: diameter, height, weight for each. 3. Inspection speed: determined by your production capacity/line speed.    Inspection Theory    Inspection process of Keye Tech vision inspection system to detect appearance defects: Feeding device: Products transfer to a workstation via feeding devices, entering the inspection machine orderly one by one. Image acquisition: High-pixel images are captured by customized system of self-developed industrial cameras and LED light sources. Image processing: Captured images upload to our self-developed KVIS COULD AI Platform for few image annotation and model training(within 1h). AI algorithm models: AI algorithms analyze the processed images, sort products into "NG" (defective) and "OK" (qualified), and trigger inline rejection of defective products via pneumatic devices. Software scheduling: The software provides sample comparison standards, allowing customers to select the most suitable criteria to balance accuracy and defect rates. Counting and box division: Defective products are instantly rejected into a separate box. Good products are counted for box division.

Machine Vision Inspection is a technology based on image processing that is used to automate inspection processes in production lines in various manufacturing industries. Basically, it allows a computer to inspect and evaluate product quality using industrial cameras, lights, and computer vision software.   Machine vision automates specific industrial tasks like defect detection, assembly guidance, and barcode scanning, making processes faster and more cost-effective. 2. Computer vision uses AI to analyze visual data, supporting applications like facial recognition, healthcare diagnostics, and customer behavior analysis.   When computer scientists working in computer science disciplines that involve images, such as computer graphics, image processing, and computer vision, noticed that their methods and applications increasingly overlapped, they began using the term "visual computing" to describe these fields collectively.

Overview. The visual inspection machines capture high-definition images of semiconductor products and transparent objects such as glass and perform the measurements for scratches, particulates and debris, chipping, color unevenness, dimensions of devices and so forth through the image processing technology.   These systems use optic lenses, sensors, and artificial intelligence technology to detect flaws, defects, and irregularities in your products. Essentially, vision inspection systems compare manufactured products to images of what the product should look like, identifying discrepancies between the two.   The purpose of an inspection is to identify whether work equipment can be operated, adjusted and maintained safely, with any deterioration detected and remedied before it results in a health and safety risk.   Principles of Visual Inspection. The fundamental principle of visual inspection is based on the detection of surface anomalies through direct observation. Inspectors clean the surface of the material or asset they're going to inspect to remove any contaminants such as dirt, grease, or paint that could obscure defects.