When looking for the right 3D imaging technology for a particular application, a few general details about the application and the part to be inspected need to be known. In the following, I have compiled a basic list of five technical questions that I judge essential in order to choose the right technology:
? Size: What is the desired field of view (FoV; length x width x height; preferably in mm) where the part to inspect is located? Also the height is important! What is the minimal/maximal free working distance (FWD) for the 3D camera?
? Features: Which features and/or defects should be detected or measured? Are metric numbers necessary?
? Surface: Which are the characteristics of the part’s surface? Colour, shininess, transparency, finish?
? Resolution: Which is the size of the smallest feature/defect that needs to be detected? Which resolution is desired (length/width in mm/pixel, height in mm/DN; DN: digital number or gray value)?
? Speed: Is the part moving or stationary during image acquisition? If moving, is it moving at constant speed and how fast (mm/s)? If stationary, what is the maximum time (in s) the part can rest? What is the maximum cycle time (in s/part)?
I encourage you to comment on theses questions and extend the list if necessary.
I also invite you to present your unsolved 3D machine vision application here by answering the listed questions. All 3D experts in this forum will be happy to help you find solutions that meet your application's requirements if phyically possible!
Kind regards, Tobias.
more: a) standoff distance -- how close can you get or how far away does the sensor have to be? b) re Surface -- i think the key issues are surface reflectivity and roughness especially relative to sensor wavelength -- this will cover the issues of shininess, transparency, colour etc; c) number of poses or number of sensors required to enable sensing of all relevant portions of the part within the time budget;
Hi Jim, thanks for your comments! The "standoff distance" or "free working distance" (FWD) is already part of the first question about "SIZE". As to the surface characteristics, yes one could add reflectivity (~ shininess) or roughness. But also the colour is important: for black rubber, one needs more light or exposure time than for brighter materials such as alu cast. Also the degree of transparency or opacity influences the result of the height measurement. In a semi-transparent material, the light is not only scattered on the surface but also inside the material, thus, the height information is blured; this effect can only in part be compensated with shorter wavelengths of the light source (e.g. blue or UV).
@all members: Please do not hesitate to present the technical details of your new 3D application here. We will be happy to suggest you information and hints about the most suited technlogies ...
I do have an application detecting several features on a spray nozzle you would see from your everyday cleaning supplies. The manufacturer of these spray nozzles have issues with the screens seating correctly that create the spray. So stiles the screen is way off and I can detect. My biggest concern is when there is a double screen seated. If I use 3D laser camera technology from SICK what other lighting should I worry about.? The color doesn't change it is always green.
Hi Andrew, welcome in the group!
Not being an expert in spray nozzles, I am a little puzzled about the part/feature you want to detect. As fas as I have googled, a nozzle screen normally protects the nozzle inside from getting plugged. But in your case, the screen seems to create the spray geometry at the outlet of the nozzle.
I recommend that you start a new discussion here where you describe your application in more detail; my "five questions" can provide a guideline for your description.
If you want to use laser triangulation on a small object with green color, I would recommend to use a green (532 nm) or blue (450 nm) laser line, since you get less absorption than with a red (640 nm) laser. Shorter laser wavelengths also have the advantage (i) to scatter better on the surface, (ii) to produce less speckles which are more problematic at high resolution, and (iii) to provide somewhat higher resolution due to the shorter wavelength.
Looking forward to your new discussion!
Tobias -- A very good concise list for a first estimate of feasibility.
I would add to surface characteristics gathering information on surface shape. Specifically, we would like to know if there are sharp distinct features that can create correspondence using stereo imaging. Without this question, I would infer the questions are either for structured light or for moiré imaging.
Hi Perry, indeed, this list of question is intended to gather a minimum of information about the application so that an efficient discussion about a system layout can start. But in fact, answers to these questions are needed independent from the 3D technology; it's not restricted to structured light, laser triangulation or moiré technologies.
It is correct, that information about the surface shape should be given, too, not just color and shininess. In particular steep and high slopes can cause difficulties with occlusions or surface areas which can hardly nor not be imaged.