Automatic Picture Transmission

The Automatic Picture Transmission (APT) system is an analog image transmission system developed for use on weather satellites. It was introduced in the 1960s and over four decades has provided image data to relatively low-cost user stations at locations in most countries of the world. A user station anywhere in the world can receive local data at least twice a day from each satellite as it passes nearly overhead.

The broadcast transmission is composed of two image channels, telemetry information, and synchronization data, with the image channels typically referred to as Video A and Video B. All this data is transmitted as a horizontal scan line. A complete line is 2080 pixels long, with each image using 909 pixels and the remainder going to the telemetry and synchronization. Lines are transmitted at 2 per second, which equates to a 4160 words per second, or 4160 baud.

On NOAA POES system satellites, the two images are 4 km / pixel smoothed 8-bit images derived from two channels of the advanced very-high-resolution radiometer (AVHRR) sensor. The images are corrected for nearly constant geometric resolution prior to being broadcast; as such, the images are free of distortion caused by the curvature of the Earth.

Of the two images, one is typically long-wave infrared (10.8 micrometers) with the second switching between near-visible (0.86 micrometers) and mid-wave infrared (3.75 micrometers) depending on whether the ground is illuminated by sunlight. However, NOAA can configure the satellite to transmit any two of the AVHRR's image channels.

Included in the transmission are a series of synchronization pulses, minute markers, and telemetry information.

The synchronization information, transmitted at the start of each video channel, allows the receiving software to align its sampling with the baud rate of the signal, which can vary slightly over time. The minute markers are four lines of alternating black then white lines which repeat every 60 seconds (120 lines).

The telemetry section is composed of sixteen blocks, each 8 lines long, which are used as reference values to decode the image channels. The first eight blocks, called "wedges," begin at 1/8 max intensity and successively increase by 1/8 to full intensity in the eighth wedge, with the ninth being zero intensity. Blocks ten through fifteen each encode a calibration value for the sensor. The sixteenth block identifies which sensor channel was used for the preceding image channel by matching the intensity of one of the wedges one through six. Video channel A typically matches either wedge two or three, channel B matches wedge four.

...
Wikipedia