This subroutine accesses an input image for reading. It returns the size of the image and a pointer to its data.

The [n] and [x] parts of the name are optional. If used [n] indicates the number of dimensions of the data (1 to 3, the default is 2) and [x] the data type (one of R, D, I, W, UW, B or UB, the default is R). So for instance if you wanted to use 1-D data with a data type of DOUBLE PRECISION, the subroutine that you should call is IMG_IN1D.

CALL IMG_IN( PARAM, NX, NY, IP, STATUS )
CALL IMG_IN1[x]( PARAM, NX, IP, STATUS )
CALL IMG_IN2[x]( PARAM, NX, NY, IP, STATUS )
CALL IMG_IN3[x]( PARAM, NX, NY, NZ, IP, STATUS )

PARAM = CHARACTER $\ast$ ( $\ast$ ) (Given)

NX = INTEGER (Returned)

NY = INTEGER (Returned)

NZ = INTEGER (Returned)

IP = INTEGER (Returned)

STATUS = INTEGER (Given and Returned)

• It is possible to access more than one image per call using multiple parameter names. Multiple parameter names are specified by supplying a comma separated list of names (i.e. ’DATA,BIAS,FLAT’). A pointer to the data of each image is then returned (in this case the IP argument should be passed as an array of size at least the number of parameter names). The advantage of obtaining a sequence of images in this manner is that the images are guaranteed to have the same shape and the same data type.

• The message which complains when "bad" (undefined) pixels are present in the input data can be stopped by following each parameter name by the character "!", i.e. ’DATA!,BIAS!,FLAT’ will inhibit checking the images associated with parameters ’DATA’ and ’BIAS’, but will check the image ’FLAT’. "Bad" pixels have the symbolic values VAL__BAD[x] which are defined in the include file PRM_PAR.