Recently, the deep neural network (DNN) has become one of the most advanced and powerful methods used in classification tasks. However, the cost of DNN models is sometimes considerable due to the huge sets of parameters. Therefore, it is necessary to compress these models in order to reduce the parameters in weight matrices and decrease computational consumption, while maintaining the same level of accuracy. In this paper, in order to deal with the compression problem, we first combine the loss function and the compression cost function into a joint function, and optimize it as an optimization framework. Then we combine the CUR decomposition method with this joint optimization framework to obtain the low-rank approximation matrices. Finally, we narrow the gap between the weight matrices and the low-rank approximations to compress the DNN models on the image classification task. In this algorithm, we not only solve the optimal ranks by enumeration, but also obtain the compression result with low-rank characteristics iteratively. Experiments were carried out on three public datasets under classification tasks. Comparisons with baselines and current state-of-the-art results can conclude that our proposed low-rank joint optimization compression algorithm can achieve higher accuracy and compression ratios.

Acoustic detection technology is a new method for early monitoring of wood-boring pests, and the effective denoising methods are the premise of acoustic detection in forests. This paper used sensors to record Semanotus bifasciatus larval feeding sounds and various environmental noises, and two kinds of sounds were mixed to obtain the noisy feeding sounds with controllable noise intensity. Then, the time domain denoising models and frequency domain denoising models were designed, and the denoising effects were compared using the metrics of a signal-to-noise ratio (SNR), a segment signal-noise ratio (SegSNR), and log spectral distance (LSD). In the experiments, the average SNR increment could achieve 17.53 dB and 11.10 dB using the in the test data using the time domain features and frequency domain features, respectively. The average SegSNR increment achieved 18.59 dB and 12.04 dB, respectively, and the average LSD between pure feeding sounds and denoised feeding sounds were 0.85 dB and 0.84 dB, respectively. The experimental results demonstrated that the denoising models based on artificial intelligence were effective methods for S. bifasciatus larval feeding sounds, and the overall denoising effect was more significant, especially at low SNRs. In view of that, the denoising models using time domain features were more suitable for the forest area and quarantine environment with complex noise types and large noise interference.

Recording vibration signals induced by larvae activity in the trunk has proven to be an efficient method for detecting trunk-boring insects. However, the accuracy of the detection is often limited because the signals collected in real-world environments are heavily disrupted by environmental noises. To deal with this problem, we propose a deep-learning-based model that enhances trunk-boring vibration signals, incorporating an attention mechanism to optimize its performance. The training data utilized in this research consist of the boring vibrations of Agrilus planipennis larvae recorded within trunk sections, as well as various environmental noises that are typical of the natural habitats of trees. We mixed them at different signal-to-noise ratios (SNRs) to simulate the realistically collected sounds. The SNR of the enhanced boring vibrations can reach up to 17.84 dB after being enhanced by our model, and this model can restore the details of the vibration signals remarkably. Consequently, our model’s enhancement procedure led to a significant increase in accuracy for VGG16, a commonly used classification model. All results demonstrate the effectiveness of our approach for enhancing the detection of larvae using boring vibration signals.

The larvae of some trunk-boring beetles barely leave traces on the outside of trunks when feeding within, rendering the detection of them rather difficult. One approach to solving this problem involves the use of a probe to pick up boring vibrations inside the trunk and distinguish larvae activity according to the vibrations. Clean boring vibration signals without noise are critical for accurate judgement. Unfortunately, these environments are filled with natural or artificial noise. To address this issue, we constructed a boring vibration enhancement model named VibDenoiser, which makes a significant contribution to this rarely studied domain. This model is built using the technology of deep learning-based speech enhancement. It consists of convolutional encoder and decoder layers with skip connections, and two layers of SRU++ for sequence modeling. The dataset constructed for study is made up of boring vibrations of Agrilus planipennis Fairmaire, 1888 (Coleoptera: Buprestidae) and environmental noise. Our VibDenoiser achieves an improvement of 18.57 in SNR, and it runs in real-time on a laptop CPU. The accuracy of the four classification models increased by a large margin using vibration clips enhanced by our model. The results demonstrate the great enhancement performance of our model, and the contribution of our work to better boring vibration detection.

The trunk borer is a great danger to forests because of its strong concealment, long lag and great destructiveness. In order to improve the early monitoring ability of trunk borers, the representative Agrilus planipennis Fairmaire was selected as the research object. The convolutional neural network named TrunkNet was designed to identify the activity sounds of Agrilus planipennis Fairmaire larvae. The activity sounds were recorded as vibration signals in audio form. The detector was used to collect the activity sounds of Agrilus planipennis Fairmaire larvae in the wood segments and some typical outdoor noise. The vibration signal pulse duration is short, random and high energy. TrunkNet was designed to train and identify vibration signals of Agrilus planipennis Fairmaire. Over the course of the experiment, the test accuracy of TrunkNet was 96.89%, while MobileNet_V2, ResNet18 and VGGish showed 84.27%, 79.37% and 70.85% accuracy, respectively. TrunkNet based on the convolutional neural network can provide technical support for the automatic monitoring and early warning of the stealthy tree trunk borers. The work of this study is limited to a single pest. The experiment will further focus on the applicability of the network to other pests in the future.

Wood-boring pests are difficult to monitor due to their concealed lifestyle. To effectively control these wood-boring pests, it is first necessary to efficiently and accurately detect their presence and identify their species, which requires addressing the limitations of traditional monitoring methods. This paper proposes a deep learning-based model called BorerNet, which incorporates an attention mechanism to accurately identify wood-boring pests using the limited vibration signals generated by feeding larvae. Acoustic sensors can be used to collect boring vibration signals from the larvae of the emerald ash borer (EAB), Agrilus planipennis Fairmaire, 1888 (Coleoptera: Buprestidae), and the small carpenter moth (SCM), Streltzoviella insularis Staudinger, 1892 (Lepidoptera: Cossidae). After preprocessing steps such as clipping and segmentation, Mel-frequency cepstral coefficients (MFCCs) are extracted as inputs for the BorerNet model, with noisy signals from real environments used as the test set. BorerNet learns from the input features and outputs identification results. The research findings demonstrate that BorerNet achieves an identification accuracy of 96.67% and exhibits strong robustness and generalization capabilities. Compared to traditional methods, this approach offers significant advantages in terms of automation, recognition efficiency, and cost-effectiveness. It enables the early detection and treatment of pest infestations and allows for the development of targeted control strategies for different pests. This introduces innovative technology into the field of tree health monitoring, enhancing the ability to detect wood-boring pests early and making a substantial contribution to forestry-related research and practical applications.