SiftAI® Sorting, Sizing, & Grading System
Revolutionizing Sorting, Sizing & Grading With Advanced Artificial Intelligence
Leveraging artificial intelligence, our SiftAI® vision system sorts out defects in your products and sizes & sorts them with industry-leading accuracy. Our innovative technology enables your processing facility to control the quality and consistency of your products while reducing labor costs.
SiftAI® is faster, more reliable, and more cost-effective than human graders and other sorting systems. If your packing shed is still using outdated optical sensors, SiftAI® can integrate into them as a drop-in replacement, effectively transforming and improving your operations.
Features
- Automated sorting and grading
- Unrivalled precision & accuracy
- Advanced technology accommodates natural variations
- Based on AI and Machine Learning technology
- Best model performance - regular model updating
- Outstanding data accessibility
- Small footprint
- Highly versatile and customizable
- User friendly interface
- Web-based management tools
Benefits
- Improved throughput times & waste reduction allow for maximized yield
- Grade consistency through technology
- Decreased labor costs and time constraints
- Full operator visibility
- Improved efficiencies and best practices
- Handles products with mass variation while accommodating products with natural variations with industry-leading precision
- SiftAI’s software is updated frequently by a team of engineers to ensure the highest standard of accuracy is being maintained
- Our system provides continually-updated data about the products that are evaluated by SiftAI's cameras (such as foreign objects logs, defect logs, throughput metrics, and size distributions)
- The SiftAI® technology is small and is easily installed above your existing conveyor or rejection system
Applications
Typical Applications
- Potatoes (all types)
- Fruits & Vegetables
- Meat
- And more - talk to us about your applications
For detailed information about how SiftAI® can assist with food safety and quality control of specific food products, visit the individual application pages, starting here: Food & Beverage Segments.
Common System Configurations
Communication Throughout the Process
Vision-Based Lane Balancing
How It Works
KPM makes use of the latest and greatest technologies including Deep Learning. From years of experience, our engineers are able to find a solution to fit even the most complex and unique challenges faced by customers.
Artificial Intelligence & Machine Learning
Automated Optical Inspection & Sorting
Regardless of the product, SiftAI® can provide tools to inspect and sort it. Capturing reliable information on quality, foreign material, size, texture, color or whatever else is needed, can be automated. Control systems can use this information to make decisions on how to handle the product. While many competitors look to uniformity, KPM prides itself on the ability to handle the organic variation in the natural world.
SiftAI® for Potatoes
The SiftAI® system, coupled with existing carriers, reliably scans potatoes for defects while assessing size and grade. Our state-of-the art hardware and AI software allows the shed operator to specify what percentage of each defect is acceptable. When defects are identified, SiftAI ejects them with precision. This system also gathers accurate size and grade estimates, sending potatoes into the designated drop.
SiftAI® for Meat
As meat cuts run under the SiftAI® system, at high speeds, the system reliably determines the size and grade of each piece of meat. SiftAI® sends digital outputs to a robot or PLC for each cut of meat that designates a location according to its classification (size, shape, color, grade, etc,).
SiftAI® for Root Vegetables
The SiftAI® system, coupled with existing carriers, reliably scans root vegetables for defects while assessing size and grade. Our state-of-the art hardware and AI software allows the operator to specify what percentage of each defect is acceptable. When defects are identified, SiftAI® ejects them with precision. This system also gathers accurate size and grade estimates, sending the vegetables into the appropriate drop.
Deep Learning Enabled
SiftAI® uses state-of-the-art artificial intelligence (AI) to segment, size, and grade products. The advanced platform will analyze organic variations and automatically route products based on user-defined attributes.
Automated Ejection or Diverting of Products that Don't Meet the Specifications
The automatic ejection process is a crucial feature that ensures the removal of faulty products from the production line without manual intervention. When the system detects an abnormality or defect during the inspection process, it triggers an automated mechanism that swiftly separates and ejects or diverts the affected items. Many types of ejection mechanism can be integrated: air jets, drop or lift nosebars, retracting nosebars, sweep arms, pushers, etc. To date, SiftAI ®systems have been integrated into Hagan, Exeter, and AgRay singulation and ejection systems.
Our Sorting Technology
SiftAI® systems are designed specifically for sorting and inspection. Leveraging artificial intelligence, our innovative technology enables your processing facility to control the quality and consistency of your products while reducing labor costs.
SiftAI® is faster, more reliable, and more cost-effective than human graders and other sorting systems. If your packing shed is still using outdated optical sensors, SiftAI® can integrate into them as a drop-in replacement, effectively transforming and improving your operations.
Identifiable Defects
Common Defects Detected by SiftAI®
Most vision systems can’t handle products with mass variations. SiftAI's advanced technology accounts for natural variations in your products with unrivalled precision, effectively discerning between defected and viable products. Defects in potatoes, for example, that we can find include:
- Green
- Growth Crack
- Misshapen (banana, heart, knob, pear, and other)
- Mud
- Sprouts
- New Bruise
- Old Bruise
- Pressure Bruise
- Rhizoct
- Rot
- Frozen
- And more...
Industry-leading Accuracy
AI sorting models across the potato packing industry vary in complexity, precision, and reliability. Here is a ranking of all those models and their capabilities from most to least effective:
Most vision systems can’t handle products with mass variations. SiftAI's advanced technology accounts for natural variations in your products with unrivalled precision, effectively discerning between defected and viable products. Defects in potatoes, for example, that we can find include:
- Green
- Growth Crack
- Misshapen (banana, heart, knob, pear, and other)
- Mud
- Sprouts
- New Bruise
- Old Bruise
- Pressure Bruise
- Rhizoct
- Rot
- Frozen
- And more...
Software
Operator Interface
From the at-line Human Machine Interface (HMI), the operator can monitor the inspection in real time. The system saves images of identified defects which the operator can review in detail and label for training and analytics.
Lanes Summary
From the Analytics menu, operators can find out quickly how the system is performing and react to issues in real time.
Defect - Old Bruise
Take a closer look at the defects to get a better idea of the issues within a particular batch. From here you can see how the system came to the conclusion it did.
Defect - Green
Take a closer look at the defects to get a better idea of the issues within a particular batch. Here you can see that the system correctly identified that 25% of the potato surface was green and designated it for the process stream.
Count Totals By Drop
Monitor the counts for each drop to make decisions for downstream processes.
Ejection Misses - Highest
High number of ejection misses are a signal that there could be a mechanical problem with the ejection system. Review the ejections by highest misses, total ejections, or ejections by hour simply by selecting the preferred tab.
Weight Totals
Here is where operators can keep an eye on the weight totals of the defects identified. Switch tabs to review weight totals by drop, size class, defects, classification, and totals by hour.
Specifications
Resources
We've highlighted a few knowledge center resources here to help you understand the product better. Please also visit our Knowledge Center for additional information and downloads, and don't forget that our team is always happy to answer any questions you have.
