Object Discrimination Device Using Total Pixel Value Array
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for computing object boundaries in an image of aligned objects require significant computational resources due to the two-dimensional data processing involved in protruding area extraction and separation.
Innovation Solution
An object discrimination device and method that acquire an image of aligned objects, generate a binarized image, compute total pixel values aligned orthogonally to the object direction, and use these values to determine object boundaries with reduced computational effort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If protruding area extraction and separation process is performed on two-dimensional binarized image data, then object boundaries can be computed, but the computation amount becomes large
Solution Approach 1:
The patent transforms the two-dimensional binarized image data into one-dimensional total pixel value data by summing pixel values along rows or columns. This dimensional reduction simplifies the computation from 2D array processing to 1D array processing, significantly decreasing the computation amount while preserving the essential information needed for object boundary detection
Solution Approach 2:
The patent extracts only the necessary information (total pixel values along projection lines) from the complete two-dimensional binarized image, discarding redundant spatial information. This selective extraction maintains the ability to detect object boundaries while reducing the data volume requiring processing
Data Source
AI summary
An object discrimination device includes an acquisition unit, a binarization unit, a feature extraction unit, and an area computation unit. The acquisition unit acquires an image of an imaging area where objects are aligned in an array from a direction orthogonal to the object alignment direction. The binarization unit generates, from the image, a binarized image of an object area that is an area where the object is present. The feature extraction unit computes a total pixel value for each array of pixel values aligned in a direction orthogonal to the object alignment direction in the binarized image, and generates a total pixel value array in which the total pixel values for the respective arrays are aligned in the object alignment direction. The area computation unit computes an object boundary in the alignment direction of the objects in the binarized image on the basis of the total pixel value array.