View allAll Photos Tagged MachineVision

Pentax k20d Cosinon-S 1.4/50 f4

Textures are 3-Dimensional.

That is to say, they have depth. Consider the statue as an example. It is 3-Dimensional art. The surface may be rough, or smooth, but it has depth.

 

in the 2-Dimensional arts - such as photography, painting or drawing - the perception of depth must be created by and through the use of shadows and highlights.

 

To further illustrate that patterns do NOT have depth, consider plaids as an example. They are, at their most fundamental, lines intersecting at right angles. There is no shadow, no light, and hence, no depth.

 

And just in the case you're feeling froggy, here's some reading on the topic for your perusal:

 

fmi.uni-sofia.bg/courses/graphics/image%20processing/pape...

 

www.cse.usf.edu/~r1k/MachineVisionBook/MachineVision.file...

 

www.cfar.umd.edu/~fer/postscript/fractal_final.pdf

Pentax k20d Cosinon-S 1.4/50 f5.6

A demo of the cookie jar alarm. The red cookie jar is monitored by the computer which will emit sounds when the jars shape changes.

An experiment in found machine-vision footage: vimeo.com/36239715

An experiment in found machine-vision footage: vimeo.com/36239715

Robot Readable World, Timo Arnall, 2012.

This is a self portrait of me taken in 1982 with the digital photography technology of the time.

 

The original photograph was taken by a monochrome CCT TV camera in the lab of Edinburgh University's Artificial Intelligence Dept in Forrest Hill, the analogue output stream of which was then passed through an ADC converter about the size and weight of a modern desktop PC. The resulting image was 256 x 256 pixels of 256 grey levels.

 

I was using an ICL PERQ as a graphics processor. It couldn't display grey levels, only black and white pixels, but I constructed a 4x4 superpixel with defined tiles to give me 16 grey levels, which I used for display of the the underlying 8 bit grey level image.

 

I had to write all of the image display and editing software from scratch myself, which was done in Pascal, that being the language for which the PERQ's hardware had been optimised.

 

The purpose of all this was to develop a rapid simple vision system suitable for use in assembly robotics, but as an amusing demonstration of the more general uses to which digital photography could be put I produced this image.

 

At that time we had no printers capable of producing graphic images. For graphic output the dept used a technical Polaroid camera which had a huge lens hood which fitted a computer screen. One of my local innovations was the use of an SLR as far away as possible in a dark room, which produced screen photographs of superior quality.

 

Originally a slide taken about 1982 of the screen of an ICL PERQ using my MInolta XG2 SLR, the slide recently redigitised by being photographed by my Sony DSC-R1.

 

DSC01882

Inspection, bar code and text reading processes for vision sensors can be significantly impaired by changing ambient lighting conditions. Simultaneously, employees and technicians are often bothered by the pulsing red light of normal vision sensors. Balluff’s new BVS vision sensor with built-in infrared lighting and integrated daylight filter eliminates both of these problems.

 

By using infrared light and the BVS sensor’s integrated daylight filter, visible ambient light cannot affect image quality or stability of an inspection, greatly increasing inspection and process reliability. The BVS sensor’s light is also invisible to the human eye, removing intrusive flickering to the operator, helping prevent the need for shielding or desire to tamper with it to minimize the lights effects.

 

Features of Balluff’s infrared vision sensor:

 

- 10% higher light intensity than comparable red light sensors

- Integrated daylight filter for increased process reliability

- Invisible light does not bother employees

- Safe for the eyes, certified for CE (EN 62471:2008)

- Simple to install with fixed integrated lighting, optics, and filter

 

Balluff offers a comprehensive line of vision sensing products, including: the BVS standard vision sensor (PC configurable using Windows compatible software and multiple tools per job), the BVS advanced vision sensor (with 360°geometric pattern match tool with configurable boolean logic outputs), the BVS identification sensor (identifies 1D barcodes and 2D data matrix codes to 0.10 mm module resolution) and a large variety of lighting and accessory options.

Last night I saw LaneHawk for the first time in the wild, at the scary Food 4 Less in Hollywood. I helped invent this, but the prospect of continuing to work on it is pretty much what drove me to quit that job.

 

It's watching for stuff on the bottom of grocery carts that people have forgotten to pay for (or are hoping nobody notices down there). With its camera It identifies what they are and rings them up on the cash register. The neat thing is that it's not looking at bar codes or anything, it's just recognizing the distinctive visual features--logos, "24 PACK" vs. "12 PACK", etc. And it works even if the items are at weird angles, or partially occluded by other items.

 

I wonder what people make of the weird square and bright light down near their ankles.

 

Section 647 of the California penal code makes it a misdemeanor to “secretly videotape, photograph, or record by electronic means, another identifiable person under or through his or her clothing, for the purpose of viewing the body of, or the undergarments worn by, that other person, without the consent or knowledge of that other person, with the intent to arouse, appeal to, or gratify the lust, passions, or sexual desires of that person and invade the privacy of that other person.” Does LaneHawk have passion? Well, I wrote the software.

Our Cognex In-Sight 7050 in a temporary mount on the bridge crane. Getting ready for MCHE470 mini-project 3b.

An experiment in found machine-vision footage: vimeo.com/36239715

An experiment in found machine-vision footage: vimeo.com/36239715

An experiment in found machine-vision footage: vimeo.com/36239715

Last night I saw LaneHawk for the first time in the wild, at the scary Food 4 Less in Hollywood. I helped invent this, but the prospect of continuing to work on it is pretty much what drove me to quit that job.

 

It's watching for stuff on the bottom of grocery carts that people have forgotten to pay for (or are hoping nobody notices down there). With its camera It identifies what they are and rings them up on the cash register. The neat thing is that it's not looking at bar codes or anything, it's just recognizing the distinctive visual features--logos, "24 PACK" vs. "12 PACK", etc. And it works even if the items are at weird angles, or partially occluded by other items.

 

I wonder what people make of the weird square and bright light down near their ankles.

 

Section 647 of the California penal code makes it a misdemeanor to “secretly videotape, photograph, or record by electronic means, another identifiable person under or through his or her clothing, for the purpose of viewing the body of, or the undergarments worn by, that other person, without the consent or knowledge of that other person, with the intent to arouse, appeal to, or gratify the lust, passions, or sexual desires of that person and invade the privacy of that other person.” Does LaneHawk have passion? Well, I wrote the software.

2 4 5 6 7 ••• 10 11