Optimized Small Waterbird Detection Method Using Surveillance Videos Based on YOLOv7

Waterbird monitoring is the foundation of conservation and management strategies in almost all types of wetland ecosystems. China's improved wetland protection infrastructure, which includes remote devices for the collection of larger quantities of acoustic and visual data on wildlife species, increased the need for data filtration and analysis techniques. Object detection based on deep learning has emerged as a basic solution for big data analysis that has been tested in several application fields. However, these deep learning techniques have not yet been tested for small waterbird detection from real-time surveillance videos, which can address the challenge of waterbird monitoring in real time. We propose an improved detection method by adding an extra prediction head, SimAM attention module, and sequential frame to YOLOv7, termed as YOLOv7-waterbird, for real-time video surveillance devices to identify attention regions and perform waterbird monitoring tasks. With the Waterbird Dataset, the mean average precision (mAP) value of YOLOv7-waterbird was 67.3%, which was approximately 5% higher than that of the baseline model. Furthermore, the improved method achieved a recall of 87.9% (precision = 85%) and 79.1% for small waterbirds (defined as pixels less than 40 × 40), suggesting a better performance for small object detection than the original method. This algorithm could be used by the administration of protected areas or other groups to monitor waterbirds with higher accuracy using existing surveillance cameras and can aid in wildlife conservation to some extent.

A Feature-Trajectory-Smoothed High-Speed Model for Video Anomaly Detection

High-speed detection of abnormal frames in surveillance videos is essential for security. This paper proposes a new video anomaly-detection model, namely, feature trajectory-smoothed long short-term memory (FTS-LSTM). This model trains an LSTM autoencoder network to generate future frames on normal video streams, and uses the FTS detector and generation error (GE) detector to detect anomalies on testing video streams. FTS loss is a new indicator in the anomaly-detection area. In the training stage, the model applies a feature trajectory smoothness (FTS) loss to constrain the LSTM layer. This loss enables the LSTM layer to learn the temporal regularity of video streams more precisely. In the detection stage, the model utilizes the FTS loss and the GE loss as two detectors to detect anomalies. By cascading the FTS detector and the GE detector to detect anomalies, the model achieves a high speed and competitive anomaly-detection performance on multiple datasets.

Analysis of Pedestrian Fractures in Collisions Between Small Cars and Pedestrians Based on Surveillance Videos

OBJECTIVE

To discuss the collision relationship and the cause of the fracture caused by traffic accidents in which the front of a small car collides with the side of a pedestrian while braking.

METHODS

The surveillance videos of 42 traffic accidents involving the front of a small car colliding with the side of a pedestrian while braking were collected. By analyzing the surveillance videos and the paths, the speed of the collision, the relationship between the vehicle and the pedestrian upon collision, and the movement trajectory of the human body were clearly identified. The type and severity of the injuries were also determined through autopsy. The characteristics of the human injuries and vehicle paths were analyzed according to the collision speed (<40 km/h, 40-60 km/h, 60-90 km/h), and the correlations between the fracture and the height of the pedestrian, the height of the hood and the length of the hood were discussed.

RESULTS

When a small car hits the side of a pedestrian, the front bumper first hits the lower limbs of the pedestrian, and then, the human body falls to the side of the vehicle, causing a secondary collision with the hood and front windshield; thus, the pedestrian is thrown at a speed similar to the speed of the vehicle, finally falling to the ground and sliding forward a certain distance. (1) When V is less than 40 km/h (n = 10), the pedestrian's head did not collide with the windshield, and the fatal injuries were caused by the individual striking the ground. (2) When V is greater than 40 km/h (n = 32), the majority (97%) of cases showed collision with the windshield. (3) When 40 to 60 km/h (n = 16), the pedestrian's head collided with the windshield, which can cause fatal injuries, and pelvic fractures and rib fractures occurred in 56.25% of patients. (4) When V is less than 60 km/h (n = 26), the ratio of the height of the pedestrian to the height of the hood was significantly smaller in the pelvic fracture group than in the nonpelvic fracture group (P < 0.01). (5) When 60 to 90 km/h (n = 16), there were holes in the windshield, and the pedestrians experienced severe head injuries, with cervical spine fracture occurring in 37.5% of patients, pelvic fractures occurring in 43.75% of patients, and rib fractures occurring in 31.25% of patients.

CONCLUSIONS

When V is less than 40 km/h, the vehicle does not cause severe injuries in pedestrians; when V is greater than 40 km/h, the collisions of the pedestrian's head with the windshield lead to severe head injuries and the accident can cause severe pelvic and rib fractures; when V is greater than 60 km/h, the collisions of the pedestrian's head with the windshield can cause cervical spine fracture in addition to head injuries. The occurrence of human injuries is related to not only the vehicle speed but also factors such as the height of the pedestrian, the height of the hood and the length of the hood.

Prediction of Pedestrian Crossing Behavior Based on Surveillance Video

Prediction of pedestrian crossing behavior is an important issue faced by the realization of autonomous driving. The current research on pedestrian crossing behavior prediction is mainly based on vehicle camera. However, the sight line of vehicle camera may be blocked by other vehicles or the road environment, making it difficult to obtain key information in the scene. Pedestrian crossing behavior prediction based on surveillance video can be used in key road sections or accident-prone areas to provide supplementary information for vehicle decision-making, thereby reducing the risk of accidents. To this end, we propose a pedestrian crossing behavior prediction network for surveillance video. The network integrates pedestrian posture, local context and global context features through a new cross-stacked gated recurrence unit (GRU) structure to achieve accurate prediction of pedestrian crossing behavior. Applied onto the surveillance video dataset from the University of California, Berkeley to predict the pedestrian crossing behavior, our model achieves the best results regarding accuracy, F1 parameter, etc. In addition, we conducted experiments to study the effects of time to prediction and pedestrian speed on the prediction accuracy. This paper proves the feasibility of pedestrian crossing behavior prediction based on surveillance video. It provides a reference for the application of edge computing in the safety guarantee of automatic driving.

Deep Learning-Based Congestion Detection at Urban Intersections

In this paper, a deep learning-based traffic state discrimination method is proposed to detect traffic congestion at urban intersections. The detection algorithm includes two parts, global speed detection and a traffic state discrimination algorithm. Firstly, the region of interest (ROI) is selected as the road intersection from the input image of the You Only Look Once (YOLO) v3 object detection algorithm for vehicle target detection. The Lucas-Kanade (LK) optical flow method is employed to calculate the vehicle speed. Then, the corresponding intersection state can be obtained based on the vehicle speed and the discrimination algorithm. The detection of the vehicle takes the position information obtained by YOLOv3 as the input of the LK optical flow algorithm and forms an optical flow vector to complete the vehicle speed detection. Experimental results show that the detection algorithm can detect the vehicle speed and traffic state discrimination method can judge the traffic state accurately, which has a strong anti-interference ability and meets the practical application requirements.

Authentication of Surveillance Videos: Detecting Frame Duplication Based on Residual Frame

Nowadays, surveillance systems are used to control crimes. Therefore, the authenticity of digital video increases the accuracy of deciding to admit the digital video as legal evidence or not. Inter-frame duplication forgery is the most common type of video forgery methods. However, many existing methods have been proposed for detecting this type of forgery and these methods require high computational time and impractical. In this study, we propose an efficient inter-frame duplication detection algorithm based on standard deviation of residual frames. Standard deviation of residual frame is applied to select some frames and ignore others, which represent a static scene. Then, the entropy of discrete cosine transform coefficients is calculated for each selected residual frame to represent its discriminating feature. Duplicated frames are then detected exactly using subsequence feature analysis. The experimental results demonstrated that the proposed method is effective to identify inter-frame duplication forgery with localization and acceptable running time.