Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Segmenting Wood Rot using Computer Vision Models

Roland Kammerbauer, Thomas H. Schmitt, Tobias Bocklet

TL;DR

An AI model to detect, quantify and localize defects on wooden logs and explores, train and fine-tune the state-of-the-art InternImage and ONE-PEACE architectures for semantic segmentation.

Abstract

In the woodworking industry, a huge amount of effort has to be invested into the initial quality assessment of the raw material. In this study we present an AI model to detect, quantify and localize defects on wooden logs. This model aims to both automate the quality control process and provide a more consistent and reliable quality assessment. For this purpose a dataset of 1424 sample images of wood logs is created. A total of 5 annotators possessing different levels of expertise is involved in dataset creation. An inter-annotator agreement analysis is conducted to analyze the impact of expertise on the annotation task and to highlight subjective differences in annotator judgement. We explore, train and fine-tune the state-of-the-art InternImage and ONE-PEACE architectures for semantic segmentation. The best model created achieves an average IoU of 0.71, and shows detection and quantification capabilities close to the human annotators.

Segmenting Wood Rot using Computer Vision Models

TL;DR

An AI model to detect, quantify and localize defects on wooden logs and explores, train and fine-tune the state-of-the-art InternImage and ONE-PEACE architectures for semantic segmentation.

Abstract

In the woodworking industry, a huge amount of effort has to be invested into the initial quality assessment of the raw material. In this study we present an AI model to detect, quantify and localize defects on wooden logs. This model aims to both automate the quality control process and provide a more consistent and reliable quality assessment. For this purpose a dataset of 1424 sample images of wood logs is created. A total of 5 annotators possessing different levels of expertise is involved in dataset creation. An inter-annotator agreement analysis is conducted to analyze the impact of expertise on the annotation task and to highlight subjective differences in annotator judgement. We explore, train and fine-tune the state-of-the-art InternImage and ONE-PEACE architectures for semantic segmentation. The best model created achieves an average IoU of 0.71, and shows detection and quantification capabilities close to the human annotators.
Paper Structure (15 sections, 2 equations, 4 figures, 5 tables)

This paper contains 15 sections, 2 equations, 4 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Data
  4. Computer Vision Models
  5. Intern Image
  6. ONE-PEACE
  7. Experiments and Results
  8. Evaluation Metrics
  9. Initial Experiments
  10. Casting of Rot(maybe)
  11. Semi-automatic Ground Truth Correction
  12. Analysis of the best performing Model
  13. Inter-Annotator Agreement Analysis
  14. Conclusions
  15. Future Work

Figures (4)

  • Figure 1: A selection of log crosscut images from our dataset, exemplifying the differences in position, weather, and lighting
  • Figure 2: Micro-average confusion matrix of the InternImage-H-UperNet model on the test dataset
  • Figure 3: Histograms of the sample distribution for the respective metrics on the test set
  • Figure 4: Comparison of the ground truth annotation (left) and prediction of the best model (right) to the original image (center)