Image Expression Type

The Image Expression Language is used by the five calculator functors to evaluate algebraic and logical expressions. The complete operator reference — including full syntax tables, availability notes, null value handling, EGO Script examples, and practical examples — is in Calculate Functors — Complete Operator Documentation.

Graphical representation of the image expression editor

The editor supports a few interactions that make it easier to work with identifiers:

• Hovering the mouse over an identifier (i1, t1, v1, etc.) displays a tooltip for the functor and output port it is connected to.
• Clicking an identifier inserts it at the current cursor position in the expression.
• If the functor connected to an identifier has an alias, that alias is shown in place of the numbered identifier (i1, t1, v1, etc.), making the expression easier to read.

Note: The editor works exclusively with the verbose form. Abbreviated syntax is converted to verbose form automatically when the script is read. See Calculator functor shorthand and Conversion between forms for details.

In expressions, operands are referenced by type-specific identifiers:

• iX — a connected map, where X is an integer from 1 to 100
• tX — a connected table or lookup table, where X is an integer from 1 to 100
• vX — a connected scalar value, where X is an integer from 1 to 100

Maps are connected via Number Map hook functors, tables via Number Table, and scalar values via Number Value.

In the abbreviated (shorthand) syntax, operands are referenced directly by variable name using type sigils: # for maps, % for tables and lookup tables, $ for values — eliminating the hook block entirely. See Calculator functor shorthand for details.

During logical expression evaluation, zero is false and any non-zero value is true.

A null value in any operand poisons the entire expression result. The keyword null marks the current evaluation unit as missing — it is an abstract sentinel with no numeric value, available in all five calculator functors. This is distinct from null(iX), which returns the numeric sentinel value stored in a specific image and is only available in map calculators. Two operators interrupt null propagation: isNull() and ?. See Null Value Handling for the complete rules, exceptions specific to map calculators, and defensive patterns.

Arithmetic, logical, and conditional operators with full precedence rules, including if/then/else, boolean and/or/not, comparison operators, +, -, *, /, %, ^, and the catch-error operator ?.

→ General Operators

Mathematical functions (sqrt, sin, cos, ln, exp, abs, ceil, floor, round, max, min, clamp, clampMod, signal, pi), stochastic functions (rand, rUniform, rNormal, rPoisson), and control functions (abort, isNull(), null). All functions in this section are available in all five calculator functors.

→ General Functions

Retrieves a scalar value connected to a vX port.

→ Value Operators

Operators that work at the current cell: iX (cell value), iX[line, col] (coordinate lookup with boundary mirroring), null(iX) (numeric sentinel value of image X — distinct from the abstract null keyword), line, and column. Available in Calculate Map and Calculate Categorical Map only; line and column are also available in the two lookup table calculators with row-oriented meanings.

→ Image Operators and Functions

Single-key lookup table queries using rule-operators: exact match, lower/upper bounds, closest key, linear interpolation, key existence test, and named attribute lookup.

→ Lookup Table Operators

Multi-key table queries using a set of composite keys, returning values from a named, indexed, or default data column.

→ Table Operators

Window-based spatial aggregation: nbMin, nbMax, nbSum, nbProd, nbCount, nbAverage, nbMedian, nbMode, nbVar, nbStdDev. Available in Calculate Map and Calculate Categorical Map only.

→ Image Neighborhood Functions

Neighbourhood statistics used as indices to retrieve values from a lookup table or a second image: nbMinRef, nbMaxRef. Available in Calculate Map and Calculate Categorical Map only.

→ Image Neighborhood Indexing Functions

Kernel-weighted convolution via nbConvol. Available in Calculate Map and Calculate Categorical Map only.

→ Convolution Operator

In an EGO Script .ego file, an image expression is enclosed in square brackets [ ]. The same expression can be written in verbose form (numbered identifiers, hook block required) or abbreviated form (named operands, no hook block). The following example, evaluated for each cell of a raster map, restores noisy pixels using a growing neighbourhood median window — combining multiple maps, a value operand, a neighbourhood function, and the ? catch-error operator:

Verbose form:

if i1 = 2 then
    nbMedian(i2, v1, v1) ? i3
else
    i3

i1 = map flagging original and noisy pixels; i2 = original pixels only; i3 = current restored image; v1 = window size (grows with the current time step). If the neighbourhood median fails (e.g. all neighbours are null), ? falls back to the current restored value i3.

Abbreviated form:

if #arePixelsOriginalOrNoisyWithCategories = 2 then
    nbMedian(#originalPixelsOnlyInRestoredImage, $lK, $lK) ? #currentRestoredImage
else
    #currentRestoredImage

For the full EGO Script syntax — including the hook-block form and conversion between the two — see Calculator functor shorthand.