Data reading method, image processing module, and monitoring processing method and device

Abstract

The embodiment of the invention provides a data reading method, an image processing module, and a monitoring processing method and device, and belongs to the image detection field; the data reading method comprises the following steps: polling items in a storage module, wherein the items include: whether an image data request is distributed or not, receive a channel number of an image data channel, the data bulk of the image data, and stores the image data request in a cycle queue; extracting data from the channel queue according to the contents contained by the items; sending the extracted data to corresponding channels through a shared system data channel, wherein the channel queue comprises data buffer zones of the channel, and one channel is matched with a plurality of image data requests. The image processing module comprises the cycle queue, the storage module, and a polling logic module. The method and device can realize fair dispensing of the channel queue, thus improving the locomotive image processing capability in the maximum level.

Description

technical field [0001] The invention relates to the field of image detection, in particular to a method for reading data, an image processing module, a method and a device for monitoring and processing. Background technique [0002] In recent years, the development of multi-channel image processing has provided a broad space for the use of parallel processing technology to improve product performance. Multi-channel image processing has high density and large capacity. Existing technologies often use multiple data paths for data transmission or multi-channel image processing to share a data buffer. How to ensure that each channel can reasonably and fairly occupy the total bandwidth of the system and realize data load balancing has become the primary problem faced in multi-channel image processing applications. [0003] In the existing multi-channel image processing system, each image processing has its own data channel, each of which undertakes the same work, so that data re...

AI Technical Summary

Problems solved by technology

This architecture is relatively simple, and each image processing is wear-balanced, but because all of them are responsible for the same work, it is bound to generate a lot of competition for shared resources. In order to deal with these concurrent and synchronous use of a large number of bus arbitrations, performance is restricted. When concurrency and When the consumption of synchronization reaches a certain level, the performance will not increase but will decrease

Images