Cost of image processing software
My cost sense after 25 years of automating visual inspections for a living. Considering 1 inspector x 3 years, the cost to automate visual inspections is 10 million at most. To make it work as a business, we need to secure about 20% profit, and we need to make it for 8 million. Outside [...].
Things just got interesting.
I have been going around to everyone, publicly announcing that we are aiming for sales of 1 billion yen in 5 years. Until now, I was afraid to step on the gas pedal because I thought "more business inquiries" = "more unrealistic inquiries" and "more orders" = "more work for me". But now, I am not so afraid to step on the gas pedal.
I did a "visual inspection" at search
It's been a while since I did a google searchfor "visual inspection". It's a great thing, isn't it? Large and small companies are all mixed up in a battle royal state. Our company's site doesn't come up even if you go to search....
Aiming for 1 billion
We have to be conscious of "continuity" as a responsibility to our customers.
FIS-200
FINALLY, THE FIS WITH DEEP LEARNING "TOO" HAS STARTED TO WORK. FIRST OF ALL, AS A RELIABLE METHOD, WE ADOPTED AN APPROACH THAT EVOLVED FROM THE CONVENTIONAL COMPARATIVE INSPECTION APPROACH: REGISTERING ONLY GOOD PRODUCTS AND USING DEEP LEARNING TO LEARN ONLY THE FLUCTUATIONS OF GOOD PRODUCTS [...].
The Business of Automated Visual Inspection
I WAS BROWSING THE NET AT SEARCHAND FOUND THAT AN OLD ACQUAINTANCE OF MINE IN THE SAME INDUSTRY IS BANKRUPT... THERE ARE OTHER AI-BASED VISUAL INSPECTION COMPANIES THAT ARE BANKRUPT AS WELL. I CAN'T HELP BUT WONDER ABOUT THE TRENDS OF OTHER COMPANIES IN THE SAME INDUSTRY. I HAVE BEEN WORKING ON THIS SUBJECT FOR OVER 20 YEARS [...].
In a place called Kyoto
If it had been in Tokyo or Osaka, we would have been involved in money games... If it had been in Nagoya, we would have had too many projects to work calmly... If it had been in Tohoku or Kyushu, it would have been a little inconvenient to deal with the whole country... [...].
Visual inspection that "goes beyond people
We visited a manufacturer of precision plastic parts to inspect the sixth unit of a visual inspection system and asked about the status of the delivered system. We are 99% satisfied with the system. I can't imagine not having it anymore. I am afraid it will break down. I am afraid it will break down. This is how visual inspection equipment should be. It is a "replacement for people" [...].
Interesting, but...
Ongoing themes: Support for systems with "two 45M CoaX cameras" Support for systems that use multiple PCs for parallel inspection to increase speed Embedding HALCON Anomaly Inspection (Deep Learning) in actual equipment Development of technology that will be key in the future. [...].
AI IS USED FOR VISUAL INSPECTION."
IT SEEMS MORE AND MORE COMPANIES ARE SAYING, "WE'LL USE AI TO INSPECT THE APPEARANCE." IT IS FINE TO TRY, BUT I THINK THE RESULTS WILL BE QUITE SEVERE. THE FUNDAMENTAL PROBLEM IS THAT "HUMAN REPLACEMENT" AUTOMATION DOES NOT PRODUCE A SIGNIFICANT RETURN ON INVESTMENT, AND THAT [...].