Today, various machine vision systems are used to monitor the production process and packaging of various products, including rebars, steel sheets, power equipment, crops, etc. Many of these machine vision systems operate based on deep learning algorithms. The development of these algorithms is of great importance. Many of the challenges in industry can be solved at much lower prices by using image processing cameras to capture and process images.

Recently, various mathematics have been proposed for deep neural networks, each of which has specific features in recognizing different objects in images. For example, the mathematics in the Faster R-CNN neural network is such that it can detect smaller objects in the image successfully. So, various pests in corn, wheat, rice, and dates can be detected by this network in food and agricultural industries. The mathematics of this network can also be used in detecting and counting steel rebars in the steel industry, detecting corona phenomena in power equipment in the power industry, oil tanks in the oil industry, ships in SAR images in the shipping industry, changes in SAR images that is useful in military industries to better control on changing the borders, etc. In contrast, the mathematics used in Yolo's deep neural network is useful for detecting larger objects in images; different mathematics have been created in different versions of this deep network. Therefore, this network is used to detect vehicles and self-driving in the automobile manufacturing industry, oil spills in the seas in the environment and related industries, face recognition in biometric identification systems, and so on. Some industries, such as steel, electricity and agriculture, are the most important and strategic industries in every country. One of the most important challenges of these machine vision systems is the ability to process in real-time, because the factory production line must be monitored non-stop and quickly. In addition, these machine vision systems must be highly accurate. To achieve high accuracy, it is necessary to train a deep neural network with a wide and appropriate dataset.

The purpose of this Special Issue is to provide different solutions in which a deep neural network can be developed with high speed and accuracy for monitoring different products to solve the challenges of different industries. All deep learning techniques are welcomed in all industries. Original research and review articles are welcome.

Potential topics include but are not limited to the following:

• Deep neural networks and deep learning: designing their architecture, mathematics, etc.
• Machine vision: designing the image processing camera, lighting, and interfaces
• Real-time monitoring: designing suitable program techniques and developing related software
• High accuracy: designing suitable algorithms
• Industrial applications: oil tank detection, steel rebar splitting and counting, steel sheet defect detection, ladle hot spot detection, pest detection in crops, ship detection, corona detection in power equipment, etc.
• Localization and classification of defects in industrial products
• Defect detection in industrial products
• Accurate measurement
• Packing steel rebars of different sizes