Binocular Stereo Vision Algorithm Research Articles

Depth hole filling and optimizing method based on binocular parallax image

AbstractEnvironment perception is one of the most vital function for autonomous robots while performing complex tasks in dynamic environment. In binocular stereo vision algorithm, the calculation of disparity image depends on matching algorithm, and the unmatched points form depth holes in the image. When encountering obstacles with less texture, such as blank walls, large area deep holes will often appear. Although the existing filtering algorithm can solve the problem of small area deep hole, it cannot be applied to large area deep hole. To solve this problem, a depth hole optimization algorithm is proposed, which detects large area depth holes from the image globally, fills the depth holes according to the disparity value distribution, and optimizes the weighted least squares filtering effect. The experimental results show that the average time of the algorithm is only 15 ms.

An Automatic Machine Vision-Guided System for the Propagation of Potato Test-Tube Plantlets

In manually propagating potato test-tube plantlets (PTTPs), the plantlet is usually grasped and cut at the node point between the cotyledon and stem, which is hardly located and is easily damaged by the gripper. Using an agricultural intelligent robot to replace manual operation will greatly improve the efficiency and quality of the propagation of PTTPs. An automatic machine vision-guided system for the propagation of PTTPs was developed and tested. In this paper, the workflow of the visual system was designed and the image acquisition device was made. Furthermore, the image processing algorithm was then integrated with the image acquisition device in order to construct an automatic PTTP propagation vision system. An image processing system for locating a node point was employed to determine a suitable operation point on the stem. A binocular stereo vision algorithm was applied to compute the 3D coordinates of node points. Finally, the kinematics equation of the three-axis parallel manipulator was established, and the three-dimensional coordinates of the nodes were transformed into the corresponding parameters X, Y, and Z of the three driving sliders of the manipulator. The experimental results indicated that the automatic vision system had a success rate of 98.4%, 0.68 s time consumed per 3 plants, and approximate 1 mm location error in locating the plantlets in an appropriate position for the medial expansion period (22 days).

Depth-based computational photography

A depth-based computational photography model is proposed for all-in-focus image capture. A decomposition function, a defocus matrix, and a depth matrix are introduced to construct the photography model. The original image acquired from a camera can be decomposed into several sub-images on the basis of depth information. The defocus matrix can be deduced from the depth matrix according to the sensor defocus geometry for a thin lens model. And the depth matrix is reconstructed using the axial binocular stereo vision algorithm. This photography model adopts an energy functional minimization method to acquire the sharpest image pieces separately. The implementation of the photography method is described in detail. Experimental results for an actual scene demonstrate that our model is effective.

Noncontact 3-D Coordinate Measurement of Cross-Cutting Feature Points on the Surface of a Large-Scale Workpiece Based on the Machine Vision Method

Three-dimensional coordinate measurement of feature points on the surface of a large-scale workpiece is important and difficult. Various relative measuring methods have been presented in recent years, and the machine vision method has been paid more attentions by researchers. The application of the machine vision method in 3-D coordinate measurement of feature points on the surface of a large-scale workpiece is discussed in this paper, and an accurate, simple, and new measuring method is proposed. The design of the measuring system mainly considers the following aspects: 1) the principle and composition of the measuring system; 2) the monocular vision algorithm for camera locating; 3) the calibration algorithm of the charge-coupled device (CCD) camera; 4) the image processing algorithm of cross-cutting feature points and the calculation of their 2-D image coordinates; and 5) the binocular stereo vision algorithm for depth measurement based on the large-scale coordinate measuring machine. The experimental results indicate the correctness and reliability of the new measuring method, and we believe that it will be a reliable and efficient technique for the noncontact 3-D coordinate measurement of cross-cutting feature points on the surface of a large-scale workpiece.

An automatic machine vision-guided grasping system for Phalaenopsis tissue culture plantlets

A vision-guided grasping system for Phalaenopsis tissue culture plantlets (PTCPs) was developed and tested. In manually transplanting Phalaenopsis plantlets, the plantlet is usually grasped at the root or the stem since the leaf is fragile and is easily damaged by the gripper. An image-processing algorithm for locating a grasping point was employed to determine a suitable grasping point on the roots. A binocular stereovision algorithm was applied to compute the 3D coordinates of the grasping point. Furthermore, a grasping device with a gripper suitable for grasping the PTCP plantlet was developed and tested. Finally, the binocular vision locating algorithm was then integrated with the robotic gripper in order to construct an automatic PTCP grasping system. The experimental results indicated that the automatic grasping system had a success rate of 78.2% in grasping the plantlets in an appropriate position.