A Versatile Machine Vision Algorithm for Real-Time Counting Manually Assembled Pieces

Low-Cost Education Kit for Teaching Basic Skills for Industry 4.0 Using Deep-Learning in Quality Control Tasks

The main purposes of this paper are to offer a low-cost solution that can be used in engineering education and to address the challenges that Industry 4.0 brings with it. In recent years, there has been a great shortage of engineering experts, and therefore it is necessary to educate the next generation of experts, but the hardware and software tools needed for education are often expensive and access to them is sometimes difficult, but most importantly, they change and evolve rapidly. Therefore, the use of cheaper hardware and free software helps to create a reliable and suitable environment for the education of engineering experts. Based on the overview of related works dealing with low-cost teaching solutions, we present in this paper our own low-cost Education Kit, for which the price can be as low as approximately EUR 108 per kit, for teaching the basic skills of deep learning in quality-control tasks in inspection lines. The solution is based on Arduino, TensorFlow and Keras, a smartphone camera, and is assembled using LEGO kit. The results of the work serve as inspiration for educators and educational institutions.

A Computer Vision System for Staff Gauge in River Flood Monitoring

Rivers close to populated or strategically important areas can cause damages and safety risks to people in the event of a flood. Traditional river flood monitoring systems like radar and ultrasonic sensors may not be completely reliable and require frequent on-site human interventions for calibration. This time-consuming and resource-intensive activity has attracted the attention of many researchers looking for highly reliable camera-based solutions. In this article we propose an automatic Computer Vision solution for river’s water-level monitoring, based on the processing of staff gauge images acquired by a V-IoT device. The solution is based on two modules. The first is implemented on the edge in order to avoid power consumption due to the transmission of poor quality frames, and another is implemented on the Cloud server, where the frames acquired and sent by the V-IoT device are processed for water level extraction. The proposed system was tested on sample images relating to more than a year of acquisitions at a river site. The first module of the proposed solution achieved excellent performances in discerning bad quality frames from good quality ones. The second module achieved very good results too, especially for what it concerns night frames.

Pointer Defect Detection Based on Transfer Learning and Improved Cascade-RCNN

To meet the practical needs of detecting various defects on the pointer surface and solve the difficulty of detecting some defects on the pointer surface, this paper proposes a transfer learning and improved Cascade-RCNN deep neural network (TICNET) algorithm for detecting pointer defects. Firstly, the convolutional layers of ResNet-50 are reconstructed by deformable convolution, which enhances the learning of pointer surface defects by feature extraction network. Furthermore, the problems of missing detection caused by internal differences and weak features are effectively solved. Secondly, the idea of online hard example mining (OHEM) is used to improve the Cascade-RCNN detection network, which achieve accurate classification of defects. Finally, based on the fact that common pointer defect dataset and pointer defect dataset established in this paper have the same low-level visual characteristics. The network is pre-trained on the common defect dataset, and weights are transferred to the defect dataset established in this paper, which reduces the training difficulty caused by too few data. The experimental results show that the proposed method achieves a 0.933 detection rate and a 0.873 mean average precision when the threshold of intersection over union is 0.5, and it realizes high precision detection of pointer surface defects.