Search

This paper proposes the geometry-based polylinal routing method as a solution for motion planning in object sorting exercises in manufacturing processes. This algorithm is based on geometric properties that arise from the interaction among objects within the configuration space. The method proposed in this paper, during its experimental phase, successfully generated smooth routes with a processing time ranging from 62.5-125 ms on a computer equipped with an AMD Ryzen 7 2700X Eight-Core 3.70 GHz processor and 16 GB of RAM. When compared to the RRT algorithm, it exhibits a higher efficiency of 38% to 48%, resulting in a reduction in iterative processes and a shorter response time. Therefore, the proposed method presents a viable solution for addressing motion planning scenarios in object sorting exercises. [ABSTRACT FROM AUTHOR]

In this Article, it is proposed a computer vision system, which can detect whether a coffee fruit is suitable for production or not. In order to achieve this detection two algorithms were developed, one to classify the coffee fruit in the ripe or unripe state; and the other to detect the presence of the 'coffee berry borer'. The first one uses a Bayesian Classifier to identify the colour of the fruit, and the second algorithm searches for the holes made by the coffee berry borer on the surface of the product. Moreover, a mechanical system was designed for the transportation and separation of the coffee fruits. In the first stage, coffees are transported as pictures of them are taken. At the end of this stage, the separation mechanism alters the path of the fruit based on the result of the classifier. The system proposed obtained an effectiveness of 87%. [ABSTRACT FROM AUTHOR]

This paper presents an artificial vision application using the image processing toolbox of Matlab®. The purpose is to classify the type variety Tommy mango according to their characteristics and size. To achieve web using a camera that captures the image in a particular workspace for three mangos maximum. It also describes the programming structure of the M-file code, the segmentation and recognition methods for image processing and graphics and obtaining results that describe the process of classification through the network of choice. [ABSTRACT FROM AUTHOR]

This paper presents the integration of a mobile robot with a vision system, which allows to a mobile agent to detect and avoid obstacles in its path. The vision system has the ability to present a 3D model of the local scene as a point clouds representation. The hardware integration is formed by an RGB-D inexpensive camera, a processing unit and a robotic platform. For visual perception, an algorithm for reduction 3D to 2D information similar to the information provided by range sensors was implemented in order to reduce response time and manipulation data for navigation algorithms. The whole system was tested in a controlled environment, overcoming various tests of navigation and obstacle avoidance. This work exploring the capabilities of RGB-D sensors in mobile robotics at Colombian Pacific region and is the first step of a research proposal aims to implement a solution to the problem of Simultaneous Localization and Mapping (SLAM). [ABSTRACT FROM AUTHOR]

Companies engaged in the manufacture, marketing and maintenance of cylinders for liquefied petroleum gas in Colombia, stamped steel plates and welded to the product a unique serial code to be identified within the cylinder of the country park. Currently, the identification process is manual and checked approximately 7000 cylinders per day in a single factory. The main objective of this paper is to present a vision system that uses artificial neural networks to recognize the code. This system consists physically of a portable device that controls light environment and scene for the acquisition of images. Another component of the system is to adjust the image. The adjustment is based on median filtering, binarization, label, and segmentation, this processing allows more meaningful information and image discrimination. Finally, the intelligent component identification is performed with Hopfield neural networks and an algorithm that checks the development of image recognition. The effectiveness of the system was reported with experimental results obtained on the basis of error with a significant number of samples. [ABSTRACT FROM AUTHOR]