In the modern visible light observation system, the real-time and accuracy of detection and identification of targets is the core efficiency of the target observation system. Based on the multi-core DSP-TMS320C6678, this letter proposes a parallel processing algorithm for multi-core cooperative imaging target images. The algorithm takes full advantage of the multi-core processing capabilities of TMS320C6678 by combining the features of imaging target detection tasks. On the basis of comparing the impact of the three division methods (i.e., row division, column division and block division) on the processing of the imaging target image data, OpenMP is used to process the divided regions of the image in parallel. The algorithm can greatly reduce the running time of imaging target image processing by dividing the imaging target detection tasks into multiple tasks and mapping to multiple different cores. Finally, the results show that the algorithm runs stably and the parallel efficiency performance is better. Using the multi-core processing capability of C6678 and the parallelism of OpenMP, the time required for image processing of the imaging target is effectively reduced, and the real-time performance of the system observation can be satisfied.