Sample extraction method and device, computer readable storage medium and computer equipment

Abstract

The invention relates to a sample extraction method and device, a computer readable storage medium and a computer device. The method comprises the steps of obtaining a target road section in an urbanroad network; determining a target crowdsourcing trajectory from pre-stored crowdsourcing trajectories according to the target road section, wherein the target crowdsourcing trajectory is used for being input to a classification detector for basic training; obtaining the crowdsourcing trajectory of a random road section in the urban road network, and inputting the crowdsourcing trajectory of the random road section into the classification detector after basic training to obtain a machine recognition result; determining a target extraction sample in the crowdsourcing trajectory of the random road section according to the machine identification result and a truth value result obtained after the at least one labeling person performs element category labeling on the crowdsourcing trajectory ofthe random road section, wherein the target extraction sample is used for carrying out optimization training on the classification detector after basic training; and performing classification detection on at least one of the road and the road facility based on the classification detector after optimization training. By adopting the method, the detection and classification accuracy of roads and road facilities can be improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a sample extraction method, device, computer-readable storage medium and computer equipment. Background technique [0002] With the continuous acceleration of the urbanization process, the importance of urban infrastructure construction, especially the construction of urban traffic roads, has gradually become prominent. In the face of ever-changing urban traffic roads and road facilities, timely updating of city maps is of great importance to people's daily travel. key role. [0003] In the prior art, for the update of urban maps, it is usually to use artificially labeled samples to train the machine model, so that it can learn the ability to detect or classify roads and road facilities, so that the current urban roads can be updated by using the model detection and classification results, and then realize the Updates to the city map. [0004] However, due to the hi...

Images