<p>Optical observations play a crucial role in monitoring space debris, and long exposure large field-of-view telescopes exhibit robust detection capabilities for identifying space debris. Nevertheless, a substantial volume of data, intricate noise, nonlinearity, and target discontinuities significantly affect the observational process. To address these intricate celestial background conditions, an enhanced YOLOv8-based method for spatial debris detection is introduced into this study. Initially, a cross-scale feature fusion module is incorporated into the neck network, followed by a subtle processing step for the feature fusion component. Finally, the content-aware reassembly of features module is employed to replace the original upsampling module, which enhances the efficiency and accuracy of feature reconstruction, thereby achieving effective detection and identification of spatial debris targets. The study utilized a dataset comprising astronomical images captured by an open-source large-field-of-view optical telescope. The experimental results show that the detection accuracy and speed of the method are improved, and that they can meet the requirements of space debris detection in complex backgrounds.</p>