Several coordinate reference frames will be involved in the triangulation. Accurate and clear definition of the different frames involved is important.

Barycentric Reference Frame (BCRF)

Origin at the Solar system barycentre. Orientated to align with the ICRF (?)

Earth Centred, Earth Fixed (ECEF)

Origin at the centre of the Earth; rotates with the Earth (so non-inertial). Accurate timekeeping is required in order to transform between the rotation Earth frame and the Solar System Barycentric Frame. Note that the longitude, latitude and altitude of the camera station need to be converted to Cartesian coordinates in the Earth centred reference frame.

Conversion BCRF - ECEF

To transform positions from BCRF to ECEF it's necessary to know the position of the Earth in the BCRF as well as it's current orientation. The position is determined from the Earth's ephemeris, which is provided by the JPL Development Ephemeris and accessed via CSPICE or some other interface. The orientation is determined from the sidereal time, which can be converted from the current universal time given a suitably accurate transformation law.

Topocentric Horizon (SEZ)

Origin at the observer site; the fundamental plane is that of the local horizon, with principal axes pointing South, East and to the Zenith. Rotates with the Earth. This is the frame in which the azimuth and elevation of objects in the sky is measured.

Conversion ECEF - SEZ

The translation is determined by the observers latitude, longitude and altitude, possibly incorporating some aspherical model of the Earth topology to provide fine corrections. The rotation is determined by the latitude and longitude, again possibly incorporating corrections for the non-spherical shape of the Earth. Only the rotation is necessary for transforming the directions to reference stars. Triangulating meteors between distributed observing sites will involve the relative positions of the sites, which can either be free parameters that are solved for, or fixed parameters determined from the ECEF-SEZ transformation.

Camera frame (CAM)

3D frame with origin at the projection centre of the camera; Z axis points along the boresight (perpendicular to the detector), and the X and Y axes lie in the image frame aligned with the 'left' and 'down' directions.

Conversion SEZ - CAM

The camera frame is most intuitively specified by the azimuth and elevation of the camera boresight (the 'pointing' of the camera in the SEZ frame) and the roll angle about the boresight. The transformation can be computed by applying these rotations in the right order.

Image frame (IM)

The image frame is a 2D coordinate system with axes aligned with the rows and columns of the detector. Coordinates are denoted (i,j); the origin is at the upper left corner of the image (when looking at the image in the conventional orientation). The finite extent of the pixels must be considered, so that formally the origin is at the upper left corner of the upper left pixel, with the lower right corner of that pixel having the coordinate (1,1) and the lower right corner of the lower right pixel having the coordinate (width, height). When an event is detected in a certain pixel, most often it is considered to lie at the coordinate of the pixel centre, which ends (.5,.5)

Conversion CAM - IM

The transformation of positions in the CAM from to the IM frame is a projection defined by the projective model adopted for the camera. In general this will be a pinhole camera possibly with radial distortion.