Image Processing Apparatus for Accurate Commodity Layout Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing apparatuses face challenges in accurately checking the layout of commodities on shelves across a wide range, as they require images taken from a distant position, which can lead to incomplete or inaccurate representations, especially when the area cannot be fully photographed in a single shot.
Innovation Solution
An image processing apparatus with a photographing unit that captures images with varying photography ranges, allowing for gradual acquisition of images, and includes a commodity map generator and determination unit to compare the actual layout with planned layout information, generating a commodity inspection map to highlight differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If images are taken from a distant position to check commodities arranged in a wide range, then the coverage area increases, but the measurement precision and manufacturing precision deteriorate
Solution Approach 1:
The patent divides the wide-area layout inspection into multiple segmented images captured at different positions. The image acquiring unit sequentially captures images of different sections (e.g., left side, right side, center) rather than attempting to capture the entire area in one distant shot. This segmentation allows each image to be captured with sufficient resolution while collectively covering the complete layout area.
Solution Approach 2:
The patent transitions from a single distant 2D overview shot to multiple closer 2D images that are then synthesized into a comprehensive 3D-like spatial understanding. By capturing images from multiple positions and synthesizing them, the system achieves both wide coverage and high precision that would be impossible with a single distant image.
2Device complexity
If a one-shot image is used to check commodity layout, then the inspection process is simplified, but the measurement precision deteriorates when the area is too large
Solution Approach 1:
The patent segments the layout inspection process into multiple image capture steps rather than relying on a single one-shot image. The image acquiring unit captures multiple images of different sections sequentially, and the synthesis unit combines them into a complete layout map. This maintains relative process simplicity while dramatically improving measurement precision for large areas.
Solution Approach 2:
The patent creates a synthesized copy of the complete layout by combining multiple captured images. Rather than attempting to capture the entire layout in one image, the system creates a composite representation that preserves the precision of close-up shots while showing the complete arranged space.
3Measurement precision
If multiple images with varying photography ranges are acquired, then the measurement precision improves, but the loss of time increases
Solution Approach 1:
The patent performs preliminary actions by capturing images in a systematic sequence (e.g., left to right, or center outward) and pre-processing them during acquisition. The image acquiring unit is designed to efficiently capture multiple images with varying ranges in advance, and the synthesis unit processes them concurrently, reducing the overall time penalty compared to sequential analysis.
Data Source
AI summary
According to one embodiment, an image processing apparatus includes an image acquiring unit, a commodity identifying unit, a commodity map generator and a commodity determination unit. The image acquiring unit acquires a number of images that are photographed, with photography ranges being varied gradually. The image identifying unit identifies a commodity and the position of that commodity based on the photographed images. The commodity map generator generates a commodity map from the photographed images, based on the commodity and commodity position identified by the image identifying unit. The commodity determination unit generates a commodity inspection map which represents differences between the commodity and commodity position shown in the commodity map and those shown in commodity layout plan information representing a commodity layout plan.