Use of an HDR image in a visual inspection process
Embodiments of the invention provide a system and method for integrating HDR images in video processing in a time efficient process. In one embodiment, a method for machine vision image processing includes obtaining a plurality of images of a scene, each image having a different exposure value. The images are captured by a camera having a specific dynamic range. Pixel values of the plurality of images are compared to the dynamic range of the camera to determine, based on the comparison, a minimal number of optimal images. This minimal number of optimal images is combined, to obtain an HDR image of the scene. The scene may include an object on an inspection line.
1. A method for machine vision image processing, the method comprising:
obtaining a plurality of images of a scene from a camera having a dynamic range, each image having a different exposure value;
detecting pixels associated with an object in each of the plurality of images of the scene;
comparing pixel values of pixels associated with the object to identify a portion of each image having pixel values which are closest to a middle of the dynamic range of the camera to obtain a minimal number of optimal images;
combining the minimal number of optimal images to obtain an HDR image of the scene;
determining an exposure level Eopt which enables a highest level of information of the object;
capturing an image of the object at Eopt; and
detecting the object in the image captured at Eopt;
wherein the minimal number of optimal images combined to obtain the HDR image includes the image captured at Eopt.
2. The method of claim 1 , wherein an optimal image is determined based on a difference between values of pixels of the image to a middle value of the dynamic range.
3. The method of claim 1 , further comprising:
receiving indications of borders of the object from a user; and
determining which pixels are associated with the object based on the borders of the object.
4. The method of claim 1 , wherein the object is an area of interest within the scene.
5. The method of claim 1 , wherein the object is an item on a visual inspection line or an area of interest on the item.
6. The method of claim 1 , further comprising:
using the HDR image of the scene in an iterative process of analysis of the scene.
7. The method of claim 6 , wherein the process of analysis comprises detecting a defect on an object on a visual inspection line.
8. The method of claim 1 , wherein the identified pixel values of the optimal images are closest to a midpoint of the dynamic range of the camera.