A hybrid digital-optical joint transform correlator has been built for recognition of objects with noise and multiple objects. The correlator uses two optically addressed LCLVs as input transducer and for intensity-recording the joint transform spectrum, two optical lenses for performing a Fourier transform, a CCD camera as active output transducer, and a digital system for postprocessing the output signal and refreshing the reference image through our object library. In our experiment, the spatially separated images are imaged onto one light valve, whose output is written on the second light valve. The output of the second light valve is optically processed to produce a correlation signal at the output transducer. The correlator is supported by a digital system so that the output can be digitized for further processing. Some experimental results for multiple-object recognition are presented and discussed.