Resources

In this webinar, Andrew Bodkin shares insights into developing imaging systems that both experienced engineers and novice developers will find valuable. Over his lifelong career working with cameras, Bodkin has developed an intuitive understanding of how the elements of imaging systems interact. He shares techniques and physical interpretations to clarify the understanding of how these system elements function. This framework is invaluable in simplifying the design and build process of the optical front end of machine vision.

Andrew Bodkin, founder of Bodkin Design & Engineering, shares a technical overview of the process of machine vision sensing based on his 30 years of experience in electro-optics, custom engineering, and understanding of optical phenomenology. As the machine vision market expands and innovations become more affordable, choosing the right spectral camera or custom solution can be both exciting and overwhelming. With attention to solution-based problem-solving, Bodkin explores how to narrow in on a specific dilemma within a business’s quality control process and select the correct components for machine control.

RGB movie made from Hyperspectral images of Oxy-Propylene flame. Movies shown at 10 fps, data taken at 380-430 fps. Red: 4.4 µm; Green: 3.35 µm; Blue: 2.75 µm.

The UAV-mounted MicroCam camera developed under SBIR funding which utilizes a microbolometer-based focal plane array for capturing infrared video.

Three wavelength composite (RGB) video of a 22 wavelength hyperspectral datacube in the 2.0-5.0 μm range. The target is the hot invisible gas plume above a ½ inch high candle flame. The spatial resolution is 114 x 114 and spectral resolution is 156 nm.

Single hyperspectral datacube image from a VNIR hyperspectral video acquired at 15 frames per second is shown on the left. In this case, three wavelengths (650, 510 and 475 nm) have been assigned to the red, green and blue color channels for display. On the right are 6, out of a total of 20, single wavelength images which can be extracted from the single frame hyperspectral datacube.

The Precision Radiometric Surface Temperature (PRST) sensor delivers a surface temperature mapping capability covering a temperature range 400-2000K with resolution of <20K, with spatial resolution of 0.5 cm at a distance of 10 m, and at real-time video rates (to 132 Hz). Operating wavelength range is 2-5 μm. The key feature of this sensor is that temperature accuracy is maintained without a priori knowledge of target spectral emissivity or even as the emissivity may be rapidly changing.