The collective dynamics in neural networks is essential for information processing and has attracted much interest on the application in artificial intelligence. Synchronization is one of the most dominant phenomenon in the collective dynamics of neural network. Here, we propose to use the spiking dynamics and collective synchronization of coupled photonic spiking neurons for noisy image segmentation. Based on the synchronization mechanism and synchronization control, the noised pattern segmentation is demonstrated numerically. This work provides insight into the possible application based on the collective dynamics of large-scale photonic networks and opens a way for ultra-high speed image processing.