Zurich-based, EthonAI offers a no-code software platform for vision-based defect detection, process monitoring & root cause analysis for manufacturers. It helps them improve their financial performance (bear in mind, internal costs of quality mgmt generate on avg 15% of operating expenses) & environmental impact by reducing waste, reworks & defects as follows:
- Assembly Verification: “detect assembly errors like misplacements, wrong orientations & missing components with <99% accuracy”. The solution can be deployed in the manufacturing of PCB (to detect electrical components & their position), medical devices (to verify intactness & closure, correctness of printed labels, as well as store images to enable audit trails) & general mechanical assembly
- Surface Inspection: whether on fully automated lines (connecting to your back office) or on one’s set by operators on the factory floor (with a front-end add-on). The platform seamlessly integrates into a factory’s image acquisition set-up/ is compatible with any camera. The AI needs just 30-50 training images to calibrate. The solution can be used in the manufacturing of plastic parts (to detect surface irregularities & contaminants), packaging (to detect irregularities like stripes & bleeding) & LEDs/ displays (to detect scratches & surface cracks)
- Process Monitoring: ingesting (& sanitizing) data from multiple production lines, the AI provides multivariate statistical process controls (SPC), rendered easy-to-use via interactive visualizations (which allow one to overlay & filter by aspects such as line stops, machine status codes, product variants or defect types). A “problem solver software” (see Root Cause Analysis below) then suggest steps to remedy. The solution is currently deployed in food production (for cross-factory comparisons of quality & process parameters), pharmaceuticals (to detect temperature excursions) & plastics (to collect early warnings of anomalous processes in a unified dashboard)
- Root Cause Analysis: whether ones quality output is represented as product images (e.g. surface damages), categorical variables (e.g. pass/ fail) or numerical variables (e.g. yield), the AI identifies quality-critical parameters & their interactions, & surfaces “the non-linear & subtle dependencies that conventional methods fail to find”. The solution can be used in semiconductors (anticipated to reduce yield losses by <50%), food production & pharmaceuticals
EthonAI was founded in 2021 by Julian Senoner (CEO), Bernhard Kratzwald (CTO) & Torbjorn Netland (Advisor), as a spin-out from ETH Zurich. One year in, the startup already counts leading manufacturers like Siemens, Roche as well as Lindt & Sprüngli among its customers. (hat tip to the promise of AI in industrials). <Source: metrology.news, tech.eu, in4capital.com, eu-startups.com, techfundingnews.com>
