Recyclable material finds new ally; Vision detects bottle-crowning faults; Vision guides fluid dispensing and MORE…

Th 0512vsd Snapshots01

Recyclable material finds new ally

Defense-contractor QinetiQ (Farnsborough, UK; has developed a hyperspectral-imaging-based machine that can automatically identify and sort recyclable domestic waste. Designed to help local authorities and waste-management companies recover materials cost-effectively from curbside collections, the high-throughput demonstrator helps optimize the value and purity obtainable from recycling.

“To reclaim many of the more valuable materials from curbside collections, most recyclable waste is currently hand-sorted, which is a time-consuming, costly, and potentially dangerous,” says QinetiQ’s Stephen Takel. “By automating this process, a material-reclamation operation can run 24/7, delivering a calculated capacity of more than 36,000 tons per year.”

Th 0512vsd Snapshots01
Click here to enlarge image

QinetiQ’s is using a broadband color camera, a hyperspectral imager, a metal-detecting array, and data-fusion and classification software to identify and classify the waste items. These are then individually tracked along the conveyer until they reach the appropriate collection bin, where a series of compressed-air ejectors remove them into containers. The system is currently programmed to identify materials including Tetrapak, ferrous and nonferrous metals, and a range of plastics. The sensors could also be trained to identify other materials including glass. QinetiQ is consulting with local authorities to adopt this technology, plus exploring ways the technology could be applied in other industries with similar problems.

Vision detects bottle-crowning faults

Th 0512vsd Snapshots02v
Click here to enlarge image

Tooheys (Sydney, Australia; currently produces more than 300 million liters of beer annually from its Lidcombe brewery in Sydney. Two product lines running at 1200 bottles per minute must be frequently changed, and, since each product change requires a change in bottle crown or cap, incorrect bottle crowning can occur. Using a Cognex (Natick, MA, USA; Insight 5400, system-integrator Machinery Automation & Robotics (Silverwater, Australia; developed an inspection system that can make changes to the threshold-determining acceptance of variable product qualities-detecting scratches, for example. The camera, mounted on an adjustable post, allows for bottle-height variations while maintaining a focal length of 125 mm. The mounting bracket houses a diffuse on-axis lighting system emitting a red light; a red-cut lens cover filters out ambient light. An I/O module interfaces with the camera and the Siemens (Munich, Germany; S7 PLC, allowing product changes to be loaded into the camera.

Vision guides fluid dispensing

Th 0512vsd Snapshots03v
Click here to enlarge image

Asymtek (Carlsbad, CA, USA; designs automated fluid-dispensing systems used in the assembly and packaging of semiconductors, PCBs, FPDs, electronic components, and medical/biotech products. Two applications for Asymtek’s machines are underfill, which pulls the heat away from a semiconductor chip placed on a board, and encapsulation, which covers a chip and keeps the air off of it, leading to less corrosion and greater structural integrity. Asymtek boosted the power and usability of its dispensing equipment by incorporating a machine-vision system as a standard feature on its Century, Millennium, and new Axiom 1010 dispensing platforms. To provide enhanced pattern recognition, the vision systems include high-speed CCD cameras, LED lighting, PCVision image processors from Dalsa Coreco (St.-Laurent, QC, Canada;, and Coreco’s Sapera software to automatically correct workpiece misalignment. The systems provide enhancements to Asymtek’s pattern-recognition systems, demonstrated by an increase in speed of up to 30% for detection of standard reference marks.

Shark spotting fits a pattern

Pattern-recognition software developed by astrophysicists to locate stars and galaxies could help save the whale shark, Rhincodon typus from extinction. In the British Ecological Society Journal of Applied Ecology, Australian marine biologist Brad Norman, JAVA programmer Jason Holmberg, and astrophysicist Zaven Arzoumanian of the Universities Space Research Association (Columbia, MD, USA; write that the Groth algorithm-developed by astronomers for the comparison of two lists of coordinates such as the x, y positions of stars-could be adapted to identify individual whale sharks.

Th 0512vsd Snapshots04
Click here to enlarge image

Says Arzoumanian, “The contrast of white whale shark spots on darker skin is well suited to blob extraction. The spatial relationships between these groups form the basis for a unique identifier for each shark.” The full potential of photographic identification of whale sharks has rarely been exploited because of the unmanageable task of making visual identification in large data sets. The authors have set up the ECOCEAN Whale Shark photo-identification library to act as a repository for whale-shark photographs taken by divers and tourists, as well as researchers.

IR camera sees slag

Westar Energy (Topeka, KS, USA; needed to monitor for slag problems in boilers that drive three 800-MW generators at its St. Marys, KS, USA, electric-generating plant. Traditionally, operators would wear protective visors to see through inspection ports for signs of slag buildup. “That’s typical practice in the utility industry, and we make multiple inspections per day,” says Carl Schultz, senior PdM analyst for thermography at Westar. Westar selected the portable, dual-range MikroScan 7400 camera from Mikron Infrared (Hancock, MI, USA;, which offers three selectable termperature ranges, including a high-temperature (400°C-1600°C) range needed for IR imaging inside the boilers where temperatures can exceed 1100°C. IR filtering allows dual-spectral-band operation-8- to 14-μm long-wavelength mode or mid-wavelength with 3.9-μm microfilter for through-the-flame imaging. “This lets us use the camera for predictive maintenance monitoring in the long-wavelength band in the ambient to 400°F range,” says Schultz. “Then we can switch to the 3.9-μm band and high-temperature range to image the boiler tubes.”

More in Cameras & Accessories