flir_thermal.analysisRun.roiTemperatureMeanReading

Description

Interprets two points as opposite corners of an inclusive rectangular ROI. Bounds are rounded, ordered, and limited to the image. NaN, Inf, and -Inf pixels are excluded from the extrema and arithmetic mean. Coordinates are one-based: x is the column and y is the row. No graphics are created.

Inputs

temperatureC

Two-dimensional numeric temperature matrix in degrees Celsius. The values are converted to double before calculation.

startXY

Numeric [x y] image coordinate at one ROI corner.

endXY

Numeric [x y] image coordinate at the opposite corner. Corner order is irrelevant. Additional values are ignored; the first two values of each corner must be finite.

Outputs

hotSpot

Scalar structure with x, y, and temperatureC for the finite ROI maximum.

coldSpot

Scalar structure with x, y, and temperatureC for the finite ROI minimum. Ties select the first value in column-major order.

meanReading

Scalar structure with x and y at the clamped upper-left ROI corner, inclusive width and height, mean temperatureC, and pixelCount. pixelCount counts only finite pixels, whereas width and height describe the entire clamped rectangle.

Failure Behavior

Empty or nonmatrix temperature data, invalid corners, or an ROI with no finite pixels returns NaN point fields. meanReading contains NaN geometry and temperature and a pixelCount of zero.

Example

T = [10 NaN 30; 40 50 60; 70 80 90];
[hot, cold, avg] = ...
flir_thermal.analysisRun.roiTemperatureMeanReading(T, [1 1], [2 2]);
assert(hot.temperatureC == 50 && cold.temperatureC == 10)
assert(avg.temperatureC == 100/3 && avg.pixelCount == 3)