Target detection model automatic iteration method and device and storage medium

Abstract

The invention discloses an automatic iteration method and device for a target detection model, and a storage medium. The method comprises the steps: obtaining a target detection result through the reasoning of a vehicle-end target detection model by using vehicle-end calculation resources; according to the target detection result, collecting data valuable for performance improvement of the vehicle end target detection model; marking the data valuable for improving the performance of the vehicle end target detection model through a data driving model by using cloud calculation resources, and training the vehicle end target detection model by using a marking result; and iterating the vehicle end target detection model which is being used by the vehicle end calculation resources into the trained vehicle end target detection model. According to the method, a vehicle end reasoning-cloud training mode is adopted, valuable data are automatically collected in a targeted manner, data labeling is automatically completed, the resource advantages of the cloud are fully played, and the iteration efficiency of the automatic driving vehicle end target detection model is improved.

Description

technical field [0001] The present invention relates to the technical field of automatic driving, in particular to an automatic iteration method, equipment and storage medium of a target detection model. Background technique [0002] Autonomous driving long-tail scenarios refer to sudden, low-probability, and unpredictable scenarios, such as intersections with traffic signal failures, drunk-driving vehicles, and balloons in the middle of the road. How to deal with long-tail scenarios has always been a difficult problem in the industry, and has gradually become the key to restricting the development of autonomous driving. To solve these problems, the autonomous driving system needs to accumulate a large amount of data and continuously optimize the model. [0003] According to statistics, the number of cars in the world exceeds 1 billion, but an accident occurs every 30 seconds on average, which shows that traffic accidents are low-frequency events. In order to realize the ...

AI Technical Summary

Problems solved by technology

[0004] According to the above data, in order to fully test the safety of the autonomous driving system, at least hundreds of millions of kilometers of road tests are required, which means that tens of thousands or even hundreds of thousands of vehicles have been running 24 hours a day for hundreds of days. The efficiency of problem data is very low, resulting in higher and higher cost of model iteration and verification

Images