Image analysis for meat

A process for identifying a target section within an image composed of an array of pixels, the process including storing color data for each pixel indicative of color information at the respective part of the image, the process comprising the steps of:

• (A) discriminating from the color data the pixels having a predetermined color characteristic of the target section, and from these discriminated pixels, creating a refined data set of color data for selected pixels, the refined data set comprising data for only those selected pixels which are adjacent at least one other selected pixel and so that the refined data set encompasses a refined image section in the general target area of the image where the target section is expected to be located,
• (B) excising from the refined data set at least one data set for a cluster of pixels representing an external image section which is adjacent to and contacting the target section but which does not form part of the target section, the excising step including:
• (i) analyzing the shape of the refined image section represented by the refined data set including forming internal links between the boundary of the refined image section and individual pixels or clusters of pixels having color characteristics not matching the predetermined color characteristics of the target section, including forming both internal boundary to pixel links and also pixel to pixel links, the internal links dividing the refined image section into sub-sections, identifying valid links demarcating the target section from adjacent touching external image sections by at least one of the steps of examining properties of the links and examining properties of the sub-sections formed by the links, and retaining in the refined data set only those data representing pixels to one side of the valid links representing the target section, and
• (C) processing the refined data set after the excising step (B) has been carried out as the data set representing the target section of the image.

Apparatus for identifying a target portion within an object, the apparatus comprising means for illuminating the object, a camera directed towards the illuminated object and operative to produce color signals indicative of the colors of the illuminated object, means for receiving from the camera the color signals and for producing an image of the object composed of an array of pixels, the apparatus being operative to identify the target section of the image representing the target portion of the object, the apparatus including storage means for storing color data for each pixel indicative of color information at the respective part of the image, the apparatus comprising:

• (A) means for discriminating from the color data the pixels having a predetermined color characteristic of the target section, and from these discriminated pixels, creating a refined data set of color data for selected pixels, the refined data set comprising data for only those selected pixels which are adjacent to at least one other selected pixel and so that the refined data set encompasses a refined image section in the general target area of the image where the target section is expected to be located,
• (B) means for excising from the refined data set at least one data set for a cluster of pixels representing an external image section which is adjacent to and containing the target section but which does not form a part of the target section, the excising means including:
• (i) means for analyzing the shape of the refined image section represented by the refined data set, means for forming links between points on the boundary of the refined image section to divide the refined image section into sub-sections, means for identifying valid links demarcating the target section from adjacent touching external image sections by at least one of the steps of examining properties of the links and examining properties of the sub-sections formed by the links, and means for retaining in the refined data set only those data representing pixels to one side of the valid links representing the target section, and
• (C) means for processing the refined data set after the excising means has been operative to produce the final data set representing the target section of the image.