Efficient Single-Shot Multibox Detector for Construction Site Monitoring

Abstract

Asset monitoring in construction sites is an intricate, manually intensive task, that can highly benefit from automated solutions engineered using deep neural networks. We use Single-Shot Multibox Detector -SSD, for its fine balance between speed and accuracy, to leverage ubiquitously available images and videos from the surveillance cameras on the construction sites and automate the monitoring tasks, hence enabling project managers to better track the performance and optimize the utilization of each resource. We propose to improve the performance of SSD by clustering the predicted boxes instead of a greedy approach like non-maximum suppression. We do so using Affinity Propagation Clustering -APC to cluster the predicted boxes based on the similarity index computed using the spatial features as well as location of predicted boxes. In our attempts, we have been able to improve the mean average precision of SSD by 3.77% on custom dataset consist of images from construction sites and by 1.67% on PASCAL VOC Challenge.

Optimized YOLOv3 Algorithm and Its Application in Traffic Flow Detections

“…The "Recall" (recall rate) is the ratio of the number of correctly detected statistical vehicles to the total number of vehicles in the data set. The calculation method is shown in Equations ( 8)- (10).…”

“…x r , x r = max(Precision) (10) In these formulas, True Positive (TP) indicates the number of correctly detected vehicles, True Negative (TN) indicates the number of correctly detected backgrounds, False Positive (FP) indicates the number of incorrect detections, and False Negative (FN) indicates the number of missed detections, respectively. In Equation ( 10), x r is maximum value of the recall rate greater than the corresponding precision rate of the interval segment, and then the average value of the maximum value of 11 points is calculated.…”

“…At present, there are two main methods of target detection in deep learning: one is a target detection algorithm combining convolutional neural networks and candidate region suggestions, represented by region-based convolutional neural networks (R-CNN) [8] and spatial pyramid pooling (SPP)-net [9]. The other is to use the series of the Single Shot MultiBox Detector (SSD) [10] and the You Only Look Once (YOLO) model [11][12][13] as the representative detection algorithms to convert the target detection problem into the regression problem by machine learning. The R-CNN algorithm uses a selective search to select the region suggestion box, which improves the accuracy of target detections.…”

Using Deep Learning to Find Victims in Unknown Cluttered Urban Search and Rescue Environments

“…The most popular single-stage detectors are SSD [25], YOLOv2 [36], YOLOv3 [37•], and RetinaNet [30•]. They have been adopted for real-time object detection in self-driving cars and environment monitoring applications [38][39][40]. However, they have not yet been applied to cluttered USAR scenes.…”

3D convolutional neural network‐based one‐stage model for real‐time action detection in video of construction equipment

“…Recently, several studies that used the convolutional neural network (CNN) based on big data have been conducted in two categories (W. Fang et al, 2018;Roberts & Golparvar-Fard, 2019;Yan et al, 2020): (i) Object detection, which combined object localization and classification; and (ii) action detection, which can recognize and classify the motion of equipment or workers. The object detection model predicts the type of construction equipment and its location using the imagebased two-dimensional (2D) CNN (e.g., Faster regionbased CNN (R-CNN) and Single shot multibox detector (SSD); W. Fang et al, 2018;Thakar et al, 2018). However, 2D CNN only extracts spatial features and is inappropriate for classification or detection of the actions that require temporal features from a sequence of images.…”