so ive been doing cell tower inspections for about 8 months now and mostly been running straight RGB with a good zoom camera, works fine for visual defects, rust, loose hardware etc. but a few clients have been asking about thermal lately and im wondering if its actually useful on towers or if its more of a roofing/solar panel thing

the way i see it, thermal on a tower isnt going to tell you much that a high-res zoom wont, unless youre looking for something specific like a hot connection or overheating coax feedline. but i could be totally wrong here, i havent actually run thermal on a tower job yet. anyone doing this regularly? what sensor are you using and is the data actually useful to the client or does it just look impressive in the report

also curious about workflow ? do you fly thermal and RGB on the same flight with a dual sensor setup or are you doing separate passes. my current rig cant do both at once so id have to land and swap which is annoying on a confined site

thermal on towers is niche but not useless. we started offering it about a year ago after a carrier client specifically asked for it post-storm to check the feedline connections and antenna mounts for heat anomalies. found one actually ? a loose N-connector that was arcing slightly, showed up pretty clearly as a hot spot. wouldnt have caught it visually at all. that said its not something i pitch on every job, its more of an add-on when they have a specific concern

running a Zenmuse H20T on an M300 so we get thermal and zoom on the same gimbal which makes it way less painful. doing separate passes with a sensor swap sounds like a nightmare on a guyed tower site honestly

yeah the dual sensor thing is kind of the only way to make thermal inspection economically viable imo. if youre burning extra flight time doing a second pass the cost benefit gets weird real fast especially on a taller structure where your battery is already getting hit. i did a few tower jobs with a single thermal sensor and the clients were kind of unimpressed because they still wanted the RGB documentation anyway so it was basically double the work for not much more money