ANN ARBOR SENSOR SYSTEMS USES BLACKFIN® FOR THERMAL INFRARED IMAGING CAMERAS

Ann Arbor Sensor Systems, LLC

Ann Arbor




There's an old adage, "seeing is believing," but then, some things are just not visible to the human eye, especially under cover of darkness or fog or a blanket of thick smoke. Take for example, the dimensional accuracy required of various parts manufactured by processes such as die-casting and plastic injection molding that simply can't be inspected by sight. While Superman had X-ray vision, the rest of us are going to have to rely on thermal infrared imaging cameras. Ann Arbor Sensor Systems, LLC of Dexter, Michigan, develops and manufactures such cameras. When the company was looking for a processor to power its AXT100 thermal infrared imaging camera, it chose Blackfin® ADSP-BF533 processors from Analog Devices, Inc. (ADI). The Blackfin processors helped the company develop a camera that broke the price/performance barriers that have plagued the thermal infrared imaging camera market.

What is Die-Casting?

Breakthrough Infrared Camera
Thermal infrared imaging cameras provide visual representations of the infrared energy given off by objects. To define infrared energy, it is the heat radiated by an object, caused by the radiation resulting from the motion of the atoms and molecules; the higher the temperature, the more the atoms and molecules move, and the more infrared energy. In fact, because infrared energy gives off heat, with warmer objects "glowing" and colder objects appearing less intense, one can use thermal infrared imaging cameras to "see" in the dark, through fog, or through a blanket of smoke that visible light will not penetrate. For this reason, the cameras have often been used for surveillance, navigation, or fire-fighting applications.

Within a thermal infrared imaging camera, the lens directs infrared energy into an array of sensors that change the energy into electrical signals, which are then interpreted by a signal processor that transforms the signals into false-color images. The reason for the false-color images is simply that the human eye can't see infrared energy, so the camera assigns color scales to the degrees of temperature to differentiate the images.

AXT100 camera



Ann Arbor Sensor Systems' AXT100 camera, in particular, is designed around an uncooled thermopile sensor, which is the basis for millions of low-cost, hand-held non-contact temperature instruments and ear thermometers. The AXT100 camera actually uses an array of sensors called the Thermophile Focal Arrays (TPFAs), which are basically an array of detectors (not a single detector) that "gaze" at an image. As such, TPFAs not only enable high-resolution, they allow for much smaller, lighter, and more power-efficient cameras.

Low-Risk, Rapid Design
The Blackfin processor also enabled Ann Arbor Sensor Systems to develop its cost-effective solution quickly. "We needed a signal processor that offered a low-risk, rapid design cycle that did not require a lot of 'glue logic' to connect the processor to memory and peripherals," said David Kryskowski, Director of Ann Arbor Sensor Systems. "Blackfin was ideal in that regard." With prices starting at $12.95 for a Blackfin ADSP-BF533 processor, Ann Arbor Sensor Systems chose a processor with excellent price/performance. With processing performance of 600 MHz, Blackfin ADSP-BF533 processors are highly integrated system-on-a-chip solutions that combine industry-standard interfaces with a high-performance signal-processing core.

Infrared Cameras for Safety Inspections

Ann Arbor Sensor Systems developed its AXT100 camera for applications such as die-casting inspection, industrial-process monitoring, welding inspection, and plastic injection molding inspection. "Until now, fixed cameras for process-control applications were very expensive and not widely adopted because of the cost barrier," said Kryskowski. "We designed a camera for harsh environments that is small enough to fit in the palm of your hand, although most users mount it on a foundry wall. The cameras are used to control the cooling rate of the die, with different cameras offering varying degrees of precision, depending on their price ranges. Our cameras, which are highly cost-effective and provide an application appropriate level of precision, help companies reduce the scrap rate." (See sidebar for a description of die-casting and plastic injection molding.)

The camera can operate in real-time, or images can be captured via a digital input port and stored in internal memory (RAM) for later analysis using sophisticated software available from third-party vendors. Some of the camera's features include image processing that interpolates and smoothes a 32 x 32 image to 128 x 128 resolution. Full digital output/control is provided through the camera's RJ-45 10/100 BaseT Ethernet port, which is fully Power-Over-Ethernet IEEE 802.3af compliant. Two manual focus lens options are available: a 29o f/0.8 or a 22o f/1.0 lens, with a spectral range of 7 - 14 m. The camera provides composite video and S-Video outputs (including NTSC and PAL formats).

Extending Battery Life
Ann Arbor Sensor Systems' AXT100 camera, which is priced just under $5,000, uses a low-cost ADI CMOS op amp as a readout device for capturing the TPFA signals and transferring them to the AXT100 signal processor, a Blackfin ADSP-BF533, for further processing. CMOS op amps are perfect for low-power applications, since power dissipation in a detector's readout circuit is crucial.

Blackfin's architecture is based on a 0.13 m CMOS process, which is also ideal for portable battery-powered applications such as Ann Arbor Sensor Systems' AXT100 camera. Plus, Blackfin has a Dynamic Power Management feature, which provides the ability to vary the voltage and the frequency of operation (with features such as sleep mode) that lowers the overall power dissipation of the processor even further. For example, at 600 MHz, the Blackfin ASDP-BF533 processor dissipates 280 mW of power. However, at 300 MHz, this drops to 90 mW, and at 200 MHz, to 50 mW. By using Blackfin's on-chip power management features, customers can maximize battery life by using only as much processing power as required. The entire camera itself consumes 1.8 - 2.5W of power.

For memory and I/O, Ann Arbor Sensor Systems used Blackfin's synchronous bus for SDRAM and the asynchronous bus for flash memory and peripherals. The company also used Blackfin's parallel peripheral interface (PPI) port, which connects directly to video encoders, as well as Blackfin's other general-purpose I/O ports.

Rathsburg Associates/ Avnet, Inc.

Ann Arbor Sensor Systems, LLC

Blown away by uClinux
But Ann Arbor Sensor Systems said it also got a lot of leverage from the fact that it could develop its application using uClinux, an open-source embedded real-time operating system (RTOS), which is ported on Blackfin processors. Kryskowski said, "Having uClinux on Blackfin dramatically reduced our development time and costs." Ann Arbor Sensor Systems used a graphical user interface (GUI) and the Blackfin STAMP hardware reference design (schematics and production files available at www.blackfin.uclinux.org), after which the company wrote its own device-specific drivers and modified parameters. "We designed our camera using the Blackfin STAMP board and uClinux, which provided the software drivers we needed for the Ethernet interface, to name one example. We made some minor modifications to the prototype board, but the camera worked the first time with very few changes to the software, which was a complete shock to everyone," said Kryskowski. "That has never, ever happened. We were all blown away."

In the future, Ann Arbor Sensor Systems plans to use another member of the Blackfin family, the ADSP-BF561 processor, which features dual symmetric 600 MHz core processors, 328K bytes of on-chip memory, and two PPI ports for its next-generation thermal infrared imaging camera.

For more information about Ann Arbor Sensor Systems, Ann Arbor Sensor Systems, LLC, visit the company's website.

沪ICP备09046653号
Send Feedback X
content here.
content here.

Send Feedback

Close