Optimization of HOG Algorithm for Computer Vision in Industrial Robots

Individuals working in the contemporary industrial sector face a diverse range of obstacles, such as the occurrence of costly item misplacement and the following attempt to locate them. In order to tackle this matter, we are now broadening our research in the realm of object detection and recognition, a subject that has advantages for both the industrial and academic spheres. The proposed system utilizes the Matrix Laboratory (MATLAB) software and employs image transmission and receiving algorithms based on the Internet of Things (IoT) to carry out item identification based on pictures. In this scenario, two discrete 32-bit Advanced RISC Machine (ARM) controllers are effectively used. The experimental setup involves establishing a connection between a camera module and a controller using a camera. The captured image is then transmitted to the STM32F426zi microcontroller through the Universal Asynchronous Receiver and Transmitter (UART) interface. Subsequently, the image is forwarded to the CC3200 microcontroller via Wi-Fi, with the objective of minimizing transmission delay. Finally, the image is transmitted from the CC3200 to MATLAB using the User Datagram Protocol (UDP). The real-time acquisition and transmission of photos is facilitated by a four-wheeled robot, which is navigated along a black line using an infrared (IR) sensor.