Multi-view part data set acquisition device

Abstract

The invention discloses a multi-view part data set acquisition device. The device comprises a transmission mechanism, a camera assembly, a profile frame, a rotary part tray and a light source assembly. The rotary part tray is connected and fixed in the aluminum profile frame at the bottom through a bearing and is used for placing parts; the four light source assemblies are fixed to the bottom aluminum profile and provide a light source for an object. The camera is fixed on the transmission mechanism, and the transmission mechanism and the camera assembly cooperate to control the distance and angle between the camera and an object. Through precise linkage cooperation of the rotary tray, the transmission mechanism and the light source, more high-quality pictures can be collected more efficiently, and a better data set is established. One or more variables of illumination intensity, illumination angle, camera angle and camera distance can be controlled through a control variable method to research the influence of other variables on the recognition precision of the collected data set, and the optimal camera pose and the optimal light source condition of the collected data set are determined.

Description

technical field [0001] The invention belongs to the technical field of image acquisition, and in particular relates to a part data set acquisition device. Background technique [0002] With the development of artificial intelligence deep learning, machine vision technology is becoming more and more mature, and there are more and more applications such as object recognition, target detection, pose estimation, etc. In these applications, the image acquisition of objects is first required. Only with a high-quality data set can the confidence of the next step such as detection and pose estimation be better. Image acquisition of objects requires a large number of pictures in all directions and angles, and algorithms are used for training. For the traditional manual collection of pictures, it is difficult to control the camera position, camera angle, light source position and other factors during the collection process, resulting in low picture quality, which has a certain impact...

AI Technical Summary

Problems solved by technology

For traditional manual image collection, it is difficult to control the camera position, camera angle, light source position and other factors during the collection process, resulting in low image quality, which has a certain impact on subsequent operations

Images