In response to the susceptibility of traditional visual simultaneous localization and mapping (SLAM) to disturbances from moving objects in dynamic scenes, a visual SLAM algorithm based on object detection networks is proposed. Thisalgorithm introducesa modulefordetecting and rejectingdynamicfeaturepointsin thetrackingthread of ORB-SLAM2, thereby utilizing static feature points for pose estimation. Firstly, YOLOv7 is chosen as the backbone network for object detection, combined with GhostNet lightweight convolutional networks and convolution with SE attention mechanism (Conv_ SE) for effective environmental detection. Secondly, the detected objects are processed through classification, rejecting feature points associated with dynamic objects, and employing geometric constraints to furtheridentify and removepotentialmovingobjects. Finally, onlystaticfeaturepointsareusedforfeaturematching and pose estimation. Validation resultson theTUM datasetindicatethatcompared to ORB-SLAM2, theproposed algorithm achieves an average reduction of96.5% in the root mean square error (RMSE) of absolute trajectory error (ATE) in dynamic walk sequences and shows improvement in other dynamic sequences as well. Experimental evidence demonstrates thatthisalgorithm significantly enhancesthelocalization accuracyand robustnessofthe system in dynamic scenarios.