For the traditional human non-destructive testing and recognition meth ods,there are problems of insufficient accuracy and reliability as well as few kinds of defects to detect.To solve them,this paper proposes an aerospace com posite material defect detection algorithm incorporating frequency domain featur es.The algorithm can be divided into three main steps.Firstly,the input infor mation of the frequency domain of the image is added to the feature extraction b ackbone network which is used to improve the feature extraction effect of defect images.Secondly,a module of informational concentration is proposed in order to improve the visualization capability and detective accuracy of defects,and o n the basis of mask region-based convolutional neural network (Mask R-CNN),the segmentation mask loss function is improved.Fin a lly,combined with the cascaded neural network structure of cascade region-based convolutional neural network (Cascade R-CNN),a ne w instance segmentation network is formed.In addition,the proposed instance seg mentation network was experimentally verified in the aerospace composite materia l defect X-ray image data set,and the average accuracy of the model detection r eached 95.3%,which achieved better results than other instance segmentation alg orithms,such as Mask R-CNN and cascade mask region-based convolutional neural network (Cascade Mask R-CNN).The research result has bee n applied to the intelligent detection of several common aerospace composite mater ial defects in actual industrial production.