Real-time identification of smoke images by clustering motions on a fractal curve with a temporal embedding method

Automatic forest fire detection with CCD cameras requires a landscape image analysis in two stages: first the tracking of local dynamic envelopes of p ixels, and second the discrimination between the various natural phenomena that may cause such envelopes. For this second process, we have to deal wit h restrictive conditions: lack of spatial information, complexity of motion s, and real-time constraints on detection. We present here a fast algorithm adapted to the extraction of complex motions in small spatial envelopes, T he principle of the method is to extract local motions from cluster analysi s of points in a multidimensional temporal embedding space. We detail the f our successive steps of this method: temporal embedding of gray-levels, fra ctal indexing of points, chaining points into a linked list, and motion ext raction from point sequences of the linked list. (C) 2001 Society of Photo- Optical Instrumentation Engineers.