Correct FE Hoppy, BFL does not consider obstacles. Operator's policy and techniques will vary, but for the Companies / Aircraft types that I'm responsible for as Performance Engineer, there are 2 approaches.
Primarily, I only use BFL for the production of 'General Takeoff Charts and Tables', this takes care of the 'Field Limits', i.e. Accelerate-Stop or Accelerate-Go to the screen height (nominally 35 feet). Now, from the end of the Runway, obstacles for the 1st and 2nd segment climbs are evaluated, and the 2 associated 1st / 2nd segment climb limits found. Now we have 3 limits, namely (1) The Field Limit, (2) the 1st segment limit, and (3) the 2nd segment limit. The MBRW is then the lesser of the 3 limits found. This may lead to some pretty horrendous limits, particularly when the 1st segment obstacles are 'close in'.
Alternatively, a technique is described in the Operations Manual/s for when the 1st/2nd climb limits are too restrictive. Steadily reduce the Field length in increments, thus making the distance to the obstacle greater, and the required gradient less. Thus, whilst the Field limit is reducing, the obstacle limit is increasing. Keep on reducing the Field length until the Field Limit and the Obstacle limit are as close as possible to each other, and Voila!, you've done a reasonable job of optimising the takeoff weight. The big 'sufferer' in all of this of course is the accelerate-stop limit (because you've been reducing the Field length), which need not have been so as the Accelerate-Stop is unaffected by obstacles.
The techniques described here are crude, but effective in achieving reasonable takeoff weights whenever the RTOW is invalidated. It's a brief and incomplete description, but it will keep you safe.
Take Care,
Old Smokey