Target detection method based on multi-scale split attention mechanism

The embodiment of the invention provides a target detection method based on a multi-scale split attention mechanism, and relates to the technical field of vision. The method and the device are used for realizing multi-scale attention fusion and enhancement and improving the accuracy of target detection. The method comprises the following steps: acquiring a plurality of input scales with the same channel number, performing attention splitting operation on each scale, performing feature enhancement, acquiring independent channel feature information of each scale, fusing to obtain global channel feature information, acquiring global channel attention through attention extraction, and performing feature enhancement on the global channel feature information. And then enhancing the enhanced feature information output by the single scale, so that the single output scale not only contains the own enhanced feature information, but also contains the feature information of other associated feature scales, thereby achiev