How to use tof camera to quickly extract the outline of objects?
First, the system collects depth images through a depth camera based on ToF, structured light or binocular vision. Secondly, the depth map is preprocessed to obtain a grayscale image. The purpose of this operation is to dilute the background, strengthen the contour information of the object, and then convert it into a binary image with background separation.
Then, the connected domain is used to mark the objects in the binary image, the contours of the objects are extracted, and the depth coordinate set based on these contours is converted into a point cloud coordinate set. Through such steps, the three-dimensional image can be converted into a two-dimensional image, thereby speeding up the processing speed of the neural network. Next, let’s take a look at the hardware structure of this scheme. The image acquisition unit is connected to the grayscale conversion unit, and then the grayscale is connected to the binarization unit. Since the final object contour extraction is based on the binarized image, the object contour processing unit is also connected to the point cloud conversion unit at the same time. Connected to the binarization processing unit.
The above is the rapid extraction method of 3D object contour of tof camera. This solution solves the problems of large amount of calculation and complex algorithm in the contour extraction method in the traditional solution, greatly improves the speed of contour extraction, and has less difficulty in implementation and good extraction effect. Therefore, it can be well integrated into embedded systems with limited computing power.
