The first survey results were released in the global survey and analysis of Vision 2020. The report describes the main popular technologies and main challenges of machine vision technology in 2020 and the future.
1、 Popular technologies investigated:
We learned from the interviewed engineers that machine vision engineers are using the following popular technologies:
Deep learning, multispectral / hyperspectral, polarization imaging, embedded vision, 3D imaging and computational imaging.
The survey also highlights some of the main challenges and problems faced by today’s audience.
2、 The application items of the survey are as follows:
·Random object picking / box picking
·Random box unloading
·Defects / defect detection
·Mobile camera detection in vision system (mobile camera to check large parts or different positions inaccessible to robot)
·Imaging / inspection involves shiny, reflective and shiny metal parts
·High speed imaging
·Vision guided robot
·Multispectral / hyperspectral imaging
3、 Main challenges:
Respondents rated the difficulty of these items with a full score of 5. The following are the results of six points:
1. Imaging / inspection involves gloss: it involves the detection of shiny, reflective or gloss parts, and the application is prone to problems.
Up to 47% of the respondents said that their difficulty was 4 or 5, which was the most difficult of these challenges.
2. High speed imaging: high speed imaging is still a challenging problem, at least for some people.
Among all items, the difficulty of this option ranks second, and the proportion with a score of 4 or 5 is 36%.
3. Multispectral / hyperspectral imaging: 35% of respondents rated multispectral / hyperspectral imaging as 4 or 5. Surprisingly, 35% of respondents also gave a score of 1 or 2, indicating that respondents were very different in using these invisible imaging technologies.
4. Mobile camera detection in the visual system: on the other hand, 43% of the respondents rated the camera movement detection in the visual system as 1 or 2, indicating that this did not bring them many problems.
5. Defect / defect detection, vision guided robot, polarization imaging and embedded vision: 42% of respondents gave a score of 1 or 2 for defect / defect detection, vision guided robot, polarization imaging and embedded vision. They think this is a relatively easy technology to master.
6. Random object picking / box picking, random box unloading: for this item, only 41% of the people scored 1 or 2, indicating that the respondents did not have many challenges here.
CT: Editorial responsibility