Method for training action planning model and target searching method

Abstract

The invention relates to the technical field of target search, in particular to a method for training action planning model and target searching method, and the method comprises the steps: obtaining acurrent test image, a target object and a current step number; obtaining a predicted boundary frame and a boundary frame of the target object; determining a reward of the current action plan based onthe predicted size relationship between the bounding box and the bounding box of the target object; inputting the current test image, the target object and the current step number into an action planning model, and predicting the probability distribution of the next action and the reward corresponding to the next action; and carrying out reinforcement learning on the strategy network and the value network according to the reward planned by the current action and the probability distribution of the next action and the corresponding reward. On the basis of the predicted boundary frame and the actual boundary frame of the target object, a reward of the current action plan is determined; and reinforcement learning is carried out on the strategy network and the value network by using a prediction result, so that the strategy network and the value network are optimal, and relatively high search efficiency is achieved.

Description

technical field [0001] The invention relates to the technical field of target search, in particular to a method for training an action planning model and a target search method. Background technique [0002] The task of target search (detection and positioning) is the basic task of computer vision. Its purpose is to find out the target object in the picture according to the given picture, and give information such as its position and its bounding box. This task can be applied to smart transportation, smart security, etc. in real life. [0003] In recent years, with the development of artificial intelligence and deep learning, the traditional target search task has achieved remarkable results in terms of performance and speed. However, since the camera is installed at a fixed location, it can only passively receive visual information, which will cause the traditional object search task to be constrained by the camera setup. For example, the target is far away from the camer...

AI Technical Summary

Problems solved by technology

[0004] In view of this, an embodiment of the present invention provides a method for training an ac

Images