Interoperable vision component for object detection and 3D pose estimation for modularized robot control

Mae, Yasushi; Choi, Jaeil; Takahashi, H.; Ohara, Kenichi; Takubo, Tomohito; Arai, Tatsuo

doi:10.1016/j.mechatronics.2011.03.008

Cited by 13 publications

(11 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Evaluation results show that the proposed algorithm performs fairly well in the majority of the sceneries, although its performance degrades when it is required to detect transparent plastic objects. Our method exceeds the state of the art on Mae et al [8] for object detection, obtaining better results for objects that are small and have round surfaces, especially when the distance between the object and robot is larger than 1.2 m. In the future work, we propose to increase the database of objects and improve our object detection system using 3D information.…”

Section: Resultsmentioning

confidence: 87%

“…The proposed method differs from the method proposed by Mae et al [8] in three main respects: 1) We use for feature extraction the SURF [28] algorithm, while the Mae et al employed the Scale-invariant feature transform (SIFT) [34][35][36] for this task; 2) To find the best match for each feature we use the LSH [29,30], meanwhile Hough transform [35] was used by [8]; and 3) We use 10 different small objects as carton bottle, plastic bottle, and circular objects at a distance of 1.5 meters, while in the experiments of [8], they used six small static objects. Figure 10 shows the results obtained using the Mae et al method [8].…”

Section: Comparison With the Y Mae Et Al [8]mentioning

confidence: 79%

“…Figure 10 shows the results obtained using the Mae et al method [8]. As we can see, the object detection can be reliably obtained when they are in close range, but the success rate drops dramatically when it is relatively far.…”

Section: Comparison With the Y Mae Et Al [8]mentioning

confidence: 94%

“…A weakness shared by all of the above approaches is that they can fail when local image information is insufficient, that is, if the target is very small or highly occluded. To reduce these problems, Mae et al [8] included a local feature matching algorithm using local geometric consistency for object detection. When it is online, the system uses SIFT for scene feature extraction and compares them with those of the reference image.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Object Detection Using SURF and Superpixels

Lopez-de-la-Calleja¹,

Nagai²,

Attamimi³

et al. 2013

JSEA

View full text Add to dashboard Cite

This paper proposes a novel object detection method in which a set of local features inside the superpixels are extracted from the image under analysis acquired by a 3D visual sensor. To increase the segmentation accuracy, the proposed method firstly performs the segmentation of the image, under analysis, using the Simple Linear Iterative Clustering (SLIC) superpixels method. Next the key points inside each superpixel are estimated using the Speed-Up Robust Feature (SURF). These key points are then used to carry out the matching task for every detected keypoints of a scene inside the estimated superpixels. In addition, a probability map is introduced to describe the accuracy of the object detection results. Experimental results show that the proposed approach provides fairly good object detection and confirms the superior performance of proposed scene compared with other recently proposed methods such as the scheme proposed by Mae et al.

show abstract

Section: Resultsmentioning

confidence: 87%

Section: Comparison With the Y Mae Et Al [8]mentioning

confidence: 79%

Section: Comparison With the Y Mae Et Al [8]mentioning

confidence: 94%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Object Detection Using SURF and Superpixels

Lopez-de-la-Calleja¹,

Nagai²,

Attamimi³

et al. 2013

JSEA

View full text Add to dashboard Cite

show abstract

“…The basic idea of appearance based method is to extract feature and find feature correspondence from reference frame and current frame, and then to estimate the pose change between these two frames. In [5], [6], the authors exploit SIFT feature and give a closed-form solution for pose estimation. Unfortunately, the pose in reference frame is inaccurate or unknown in most practical applications.…”

Section: Introductionmentioning

confidence: 99%

Model based vehicle localization for urban traffic surveillance using image gradient based matching

Zheng

Peng

2012

2012 15th International IEEE Conference on Intelligent Transportation Systems

View full text Add to dashboard Cite

The matching between 3D model projection and 2D image data is a key technique for model based localization, recognition and tracking problems. Firstly, we propose a fitness function to evaluate the matching degree that uses image gradient information in the neighborhood of model projection. The weighting adjustment and the normalization for visible model projection are involved, which improves the correctness and robustness of fitness function. The fitness function is used for vehicle localization and the 3D pose is reduced to location and orientation. Then, we present a direct search optimization method with 3×3 search kernel for location estimation. The "disturbed particles" is used to avoid falling into local optimum and the coarse-to-fine optimization strategy is adopted to greatly reduce computational cost. Finally, we propose a 3D pose estimator to find location and orientation by optimizing the fitness function within orientation range. Experiments on real traffic surveillance videos reveal that the proposed optimization algorithm is effective and both fitness function and 3D pose estimator are correct and robust against clutter and occlusion.

show abstract