(AIMS 2022) SATMeas - An Object Detection and Measurement Approach: Canny Edge Detection Algorithm

Satyam Mishra (Vietnam National University - International School, Vietnam); Le Trung Thanh (Rector - Vietnam National University - International School, Vietnam)

In today's modern world, detecting objects in real-time and measuring their dimensions has become a challenge in many areas of industry. Many big warehouses, courier companies, airport containers at airports etc. can not always get precise/accurate measurement using human hands. In our research we have developed an application SATMeas to detect the object and give the measurements of the object in real-time. We have utilized the canny edge detection algorithm; we made some morphological changes to improve the algorithm to obtain the desired changes. These morphological procedures we did are combination of nonlinear procedures performed generally on the arrangement of pixels without changing their numeral values, erosion and dilation are the keys for morphological operations. We use the OpenCV function cv2.findContours to detect contours by identifying the shapes of the objects inside the edge outline. This is done after edge detection and closing any gaps between edges. We determine pixels per metric variable by relying on a reference object. The Euclidean distance between sets of center points was then determined to get the calculations. Putting it all together, we developed an application SATMeas to detect object in real time and measure it as well. In this way, the outlines of the distinctive objects within the frame were kept. The proposed method works exceptionally quick and efficiently, it gives 98~99% success in determining the measurement of the objects. Hope other scientists can add more features to it and industrialists can commercialize it for better and modern industrial use.