Method for improving environment feature recognition robustness of sweeping robot and application thereof

The invention provides a method for improving environment feature recognition robustness of a sweeping robot, a terminal and a computer readable storage medium, and solves the problem of positioning loss in the process that the sweeping robot uses a camera for positioning. According to the robustness analysis system, three characteristic parameters of detectability, depth robustness and rotation robustness are defined based on analysis of environment image characteristics; when a task is executed for the first time, many features with relatively high numerical values of detectability, depth robustness and rotation robustness can be found as reference features by analyzing image features of the surrounding environment. In the actual working process, the robot processes the environment image in real time, the environment image is compared and matched with the reference image features, the features in the environment can be rapidly recognized, the robustness of environment feature recognition of the moving sweep