3 Passive sensing systems for satellites

The first and most familiar passive sensing system is a camera. Just like the camera in a cellphone, it uses a complementary metal-oxide semiconductor (CMOS) or a charge-coupled device (CCD) to capture photons of light reflected off the target by a light source such as the sun. The photons captured by each pixel of the sensor are converted into digital information through an ADC and processed by the system’s processor to form pictures of Earth from space.

With these images, it is possible to look at weather patterns, ice coverage and the impact of natural disasters. However, the quality of the image is determined by the resolution of the sensor (the number of pixels), the dynamic range of the sensor (the number of photons that the pixel can hold), and the accuracy of the conversion of that information into a digital format. Figure 3-1 is a typical block diagram of an optical imaging payload for implementing passive sensing in a satellite.

While some image sensors integrate a data converter, others rely on the performance of external data converters such as the ADC3683-SP, which offer these features:

• Dual-channel, 18-bit resolution at up to 65MSPS enables the maximum image sensor dynamic range extraction.
• A noise spectral density of –160dBFS/Hz provides high signal-to-noise ratio for the clean images.
• <100mW of power consumption to reduce the heat generated from the electronics, which can affect sensor noise.
• 100krad/75MeV (-SP version) and 30krad/43MeV (-SEP version) enable use in any space orbit from LEO to GEO.
• 11mm-by-11mm ceramic quad flat pack (-SP version)

While you are most familiar with visible light, there are many other wavelengths of light not visible to the human eye, such as infrared and ultraviolet.

By looking at pictures from all of the spectrum of light, scientists can measure details such as the amount of pollutants in the atmosphere, the change in crop yields, geological formations, vegetation density and moisture. By exploring how these details change over time, scientists can predict what may have happened in the distant past but also estimate what could happen in the future.

It is possible to measure nonvisible light in three ways:

• Single-band imaging measures one band in the electromagnetic spectrum. For example, an infrared sensor detects infrared radiation to measure temperature changes.
• Multispectral imaging combines the images from multiple bands to sense phenomenon such as vegetation density that isn’t visible with a single band. A multispectral sensor measures three or more coarse spectral bands.
• Hyperspectral imaging captures images from very narrow slices of a certain band of light. Hyperspectral sensors can measure hundreds of narrow bands to identify features that cannot be seen with the coarse bands of multispectral imaging.

All of these imaging systems rely on sensor ICs that are sensitive to the specific bands of light being measured. It’s possible to use CMOS or CCD sensors for the visible or near-infrared spectrum, but they are not applicable for longer wavelengths of light. Indium gallium arsenide detectors can measure wavelengths from 900nm to 2500nm, making them suitable to see further in the infrared spectrum.

Prisms or gratings placed in front of an image sensor separate light into individual bands. Each pixel of the sensor in the y dimension senses a single band. The resulting two-dimensional image comprises all of the spectral information for each point across the line. It then becomes possible to examine the spectral composition of each individual pixel to look for patterns or characteristics of things like minerals, vegetation or pollution.

A sensor is just one of the components necessary to produce images, however. The output of the sensor must also be conditioned, digitized by a high-speed ADC, and then processed into a viewable format. As in the image sensor described above, the performance of the ADC is vital to the quality of images and must match the dynamic range of the sensor to get the best results. Additionally, it’s important to carefully select signal conditioning, clocking and power supplies components in order to not add additional noise to the sensor output or data converter. Low-noise components such as the TPS7H1111-SP RF low-dropout regulator introduces as little noise into the system as possible.