Most and least challenging technology issues in machine vision

In early December, Vision Systems Design published the first-of-its-kind Solutions in Vision 2020 global audience survey analysis, which identifies key design trends in machine vision for 2020 and into the future.

Survey respondents helped shed light on how much our audience uses the following hot-topic technologies: deep learning, multispectral/hyperspectral, polarization, embedded vision, 3D imaging, and computational imaging. While the article largely focused on these topics, the survey also highlights some of the major challenges and problems encountered by our audience today.

On a scale where one represents no challenges encountered and five represents the most challenges encountered, respondents were tasked with ranking the following:

• Random object picking/bin picking
• Random box depalletizing
• Defect/flaw detection
• Camera mobility in vision systems (Moving a camera to inspect different locations on large parts or beyond a robot’s reach)
• Imaging/inspection involving glossy, reflective, and shiny metal parts
• High-speed inspection
• Vision-guided robots
• Multispectral/hyperspectral imaging
• Polarization
• Embedded vision 

Here are six takeaways from the results:

1. Applications involving shiny, reflective, or glossy parts present a problem. Forty-seven percent of respondents indicated a 4 or 5 here, the highest of any other challenge by a margin.
2. High-speed imaging remains a challenging issue, at least for some. This option had the second most combined 4 and 5 responses of all, coming in at 36%. 
3. Thirty-five percent of respondents put a 4 or 5 for multispectral/hyperspectral imaging. Oddly enough though, 35% of respondents also put a 1 or 2, indicating that respondents are evenly split on the use of these non-visible imaging technologies. 
4. On the other side of things, 43% of respondents put either 1 or 2 for camera mobility in vision systems, indicating that it just doesn’t create a ton of issues for them. 
5. Four answers tied with 42% of respondents combined choosing a 1 or 2 on the scale: Vision-guided robots, polarization, and embedded vision. 
6. For random object picking/bin picking, 41% chose a 1 or 2, showing that more often than not, our audience does not have significant challenges here. 

View the full survey results article here.