Widely used as a construction material, concrete is composed of cement, water, sand and gravel. First, these materials are mixed, and then placing into a mold, and then filled and compacted with a vibrator. Next, the concrete is cured in a wet state. Each of these work steps is an important work that affects the quality of the concrete after hardening. However, the degree and the completion time of concrete compaction have been conventionally determined by visual judgment and feeling based on the experience of engineers. Such a judgment method has a risk of contributing to variations in the quality of concrete and deterioration of quality. Considering the situation where the number of engineers continues to decrease, there is a demand for consistent judgment indicators and technological development aimed at improving productivity. This paper describes a concrete compaction judgment system with deep learning which can be a substitute for conventional visual judgment, and proposes a practical system that can realize highly accurate judgment the concrete surface in each small area, based on the system with deep learning that the authors have developed so far. An evaluation experiments was performed to verify its performance, and its usefulness was confirmed.

Many camera and photo applications, for example image mosaicing, Simultaneous Localization And Mapping (SLAM), and so on, use image keypoints, descriptors, and their matching. Recently, obtaining light-field has become easier thanks to the commercialization of light-field cameras. Light-field is a record of not only light ray intensity but also directions, and multi-view images can be acquired with it. In this paper, a light-field descriptor with light ray directions using light-field is proposed. A light-field descriptor is defined as a set of feature vectors that change according to the direction of the light ray. In addition, a method to match these descriptors is proposed. As the method focuses on the direction of the light ray, it allows more accurate matching than a conventional method when the camera pose changes against the target. To confirm the effectiveness, the proposed method is evaluated by using light-field image recorded in a real scene. As a result, the proposed method is shown to improve the accuraccy over a conventional method. In particular, the closer the distance between the target and the camera, the better accuracy of the proposed method could be achieved.

In this paper, we discuss a method for automatic programming of inspection image processing. Generally, an automatic appearance inspection in a manufacturing process has been expected from a viewpoint of reliability and efficiency of the inspection. However, the automation of the appearance inspection using image processing technology is not always easy. This is because building the system requires proficiency in selecting the image processing logic and tuning the parameters. Therefore, we had proposed a method for automatic programming of inspection image processing system called "Image Processing Network Programming (IPNP)". In this study, we extend the functionality of IPNP. Concretely, that function is modifying an appearance inspection program automatically when it fails to detect a defect region. It is verified from experiments that the proposed method can modify inspection image processing programs.

The demand for non-contact inspection methods is increasing in various fields. However, there is a problem that it takes a great deal of effort for the operator to recognize the defect areas on the actual object from the measurement results displayed on the monitor. Authors have already proposed a three-dimensional shape measurement projection mapping system to reduce the problem. In this study, the prototype was applied to visualize the displacement distribution on an actual object. The measured displacement distributions images of flat and bent steel sheets due to small thermal local deformation, which was difficult to recognize with the naked eyes, were projected onto the actual object in real-time. The projected displacements in time-series was almost same the value shown by the laser displacement meter. The effectiveness of the projection of the distribution of evaluated values obtained from measured object was confirmed.

This paper targets analyzing the characteristics in the motions of the workers at each skill level in an assembly task. The analysis method quantizes the positional information of the gaze and hands on the workbench into eighteen areas and converts the positional information of into a code. Next, it calculates the occurrence frequency of pairs of codes and generates co-occurrence histograms. Finally, it analyzes relationship between the movement of the gaze and hands by using these histograms. The results of the analysis show that the non-dominant hand at elementary level stays in two areas, and the non-dominant hand at intermediate level moves to five areas. This suggests that workers move their non-dominant hand more efficiently at intermediate level than at elementary level. In addition, the results show that the gaze at intermediate level moves to seven areas, and the gaze at expert level moves to four areas. This indicates that the gaze at expert level remains at center of the workbench.

This study proposes a new pattern mapping method that enables keeping isometric and conformal conditions between 2D patterns and mapped patterns on freeform surfaces. Although the established method called parameterization is widely used in the CG field, there is no critical method in the manufacturing field. In this paper, a differential equation of conformal condition around an arbitrary point on the freeform surface is formulated, and uniform pattern mapping with less distortion on the freeform surface is enabled by combining the equation of the isometric condition proposed by the previous report. Moreover, the effectiveness of the proposed method is confirmed through some case studies.

In this study, the influence of the negative pressure generated during demolding in the injection molding process on the demolding resistance is clarified by comparing the demolding forces in the cases of demolding molded products with and without holes on the bottom. Further, as an evaluation method of the demolding resistance under vacuum negative pressure, it is proposed to use the overall energy spent in the mold release motions from start to end of the demolding process. Furthermore, a breathable mold, where the porous material is locally incorporated, was designed and manufactured to reduce the negative pressure during demolding process. Finally, the effectivity of reducing the demolding resistance by using the breathable mold was experimentally verified. The experimental results showed a higher reducing effect for demolding resistance could be obtained.

There is a close relationship between the vibration generated by cutting and the dynamics of the machining system. The dynamics of the machining system often have anisotropy depending on the machine tool structure and the machining setup. Therefore, the anisotropy of the machining system dynamics to suppress the vibration during cutting is investigated in this study. In this paper, the effect of the anisotropy of the machining system dynamics on the vibration stability in milling is investigated by time domain analysis. Cutting experiments are conducted to verify the analysis results.