LoRaWAN， as a wireless communication standard in Low Power Wide Area Network （LPWAN）， provides the support for the development of IoT （Internet of Things）. However， limited by the characteristics of incomplete orthogonality among Spreading Factor （SF） and the fact that LoRaWAN does not have a Listen-Before-Transmit （LBT） mechanism， the ALOHA-based transmission scheduling method will trigger serious channel conflicts， which reduces the scalability of LoRa （Long Range Radio） networks greatly. Therefore， in order to improve the scalability of LoRa network， Non-Persistent Carrier Sense Multiple Access （NP-CSMA） mechanism was proposed to replace the medium access control mechanism of ALOHA in LoRaWAN. The time of accessing the channel for each node with the same SF in LoRa network was coordinated by LBT， and multiple SF signals were transmitted in parallel for the transmission between different SFs， thus reducing the interference of same SF and avoiding inter-SF interference in the common channel. To analyze the impact of NP-CSMA on the scalability of LoRa networks， LoRa networks constructed by Lo RaWAN and NP-CSMA were compared by theoretical analysis and NS3 simulation. Experimental results show that NP-CSMA has 58.09% higher theoretical Packet Delivery Rate （PDR） performance than LoRaWAN under the same conditions， at a network communication load rate of 1. In terms of channel utilization， NP-CSMA increases the saturated channel utilization by 214.9% and accommodates 60.0% more nodes compared to LoRaWAN. In addition， the average latency of NP-CSMA is also shorter than that of the confirmed LoRaWAN at a network traffic load rate of less than 1.7， and the additional energy consumption to maintain the CAD （Channel Activity Detection） mode is 1.0 mJ to 1.3 mJ and 2.5 mJ to 5.1 mJ lower than the additional energy consumption required by LoRaWAN to receive confirmation messages from the gateway when spreading factor is 7 and 10. The above fully reflects that NP-CSMA can improve LoRa network scalability effectively.

Aiming at the problem of low generalization of the license plate detection model， which makes it difficult to reuse in different application scenes of smart transportation， a license plate detection algorithm in unrestricted scenes based on adaptive confidence threshold was proposed. Firstly， a multi-prediction head network model was constructed， in it， the segmentation prediction head was used to reduce the model reuse pre-processing work， the adaptive confidence threshold prediction head was used to improve the model detection ability， and the multi-scale fusion mechanism and bounding box regression prediction head were used to improve the model generalization ability. Secondly， a differentiable binary network training method was adopted to learn model parameters through differentiable binary transformation combined with the training of classification confidence and confidence threshold. Finally， the Connectivity Aware Non-Maximum Suppression （CANMS） method was used to improve the post-processing speed of license plate detection， and the lightweight network ResNet18 was introduced as the backbone network of feature extraction to reduce the model parameters and further improve the detection speed. Experimental results show that in 6 scenes with different constraints in Chinese City Parking Dataset （CCPD）， the proposed algorithm can achieve the average precision of 99.5% and the recall of 99.8%， and achieves the efficient detection rate of 70 frames per second， which are better than the performance of anchor-based algorithms such as Faster Region-Conventional Neural Network （Faster R-CNN） and Single Shot MultiBox Detector （SSD）. On the three supplementary scene test sets， the license plate detection accuracy of the proposed algorithm is higher than 90% in unrestricted scenes with different resolutions， different shooting distances， and different shooting angles of pitch. Therefore， the proposed algorithm has good detection performance and generalization ability in unrestricted scenes， and can meet the requirements of model reuse.

The current personalized recommendation platform has the characteristics of a wide range of data sources and many data types. With the data sparsity of the platform as an important reason for affecting the performance of the recommendation system, there are many challenges faced by the recommendation system:how to mine structured data and unstructured data of the platform to discover more features, improve the accuracy of recommendations in data-sparse scenarios, alleviate the cold start problem, and make recommendations interpretable. Therefore, for the personalized scenario of recommending Items for Users, the Heterogeneous Information Network (HIN) was used to build the association relationships between objects in the recommendation platform, and the Meta-Graph was used to describe the association paths between objects and calculate the User-Item similarity matrices under different paths; the FunkSVD matrix decomposition algorithm was adopted to calculate the implicit features of Users and Items, and for the unstructured data with text as an example, the Convolutional Neural Network (CNN) technology was used to mine the text features of the data; after splicing the features obtained by the two methods, a Factorization Machine (FM) incorporating historical average scores of Users and Items was used to predict Users' scores for Items. In the experiment, based on the public dataset Yelp, the proposed hybrid recommendation model, the single recommendation model based on Meta-Graph, the FM Recommendation model (FMR) and the FunkSVD based recommendation model were established and trained. Experimental results show that the proposed hybrid recommendation model has good validity and interpretability, and compared with the comparison models, the recommendation accuracy of this model has been greatly improved.

In an information explosion era, the large scale and structure complexity of datasets become problems in approximation calculation. Dynamic computing is an efficient approach to solve these problems. With the development of existing updating method applied to the dynamic approximation in multi-granular rough sets, a vector matrix based method for computing and updating approximations in Variable Precision Multi-Granulation Rough Sets (VPMGRS) was proposed. Firstly, a static algorithm for computing approximations based on vector matrix for VPMGRS was presented. Secondly, the searching area for updating approximations in VPMGRS was reconsidered, and the area was shrunk according to the properties of VPMGRS, effectively improving the time efficiency of the approximation updating algorithm. Thirdly, according to the new searching area, a vector matrix based algorithm for updating approximations in VPMGRS was proposed based on the static algorithm for computing approximations. Finally, the effectiveness of the designed algorithm was verified by experiments.

To calculate upper and lower approximations effectively and quickly under covering variation in the covering information systems, a relation matrix was defined by using the concept of characteristic function. Then the expressions for the approximations, positive, boundary and negative regions intuitively from the view of matrix were presented. Then, the expressions for the approximations, positive boundary and negative regions intuitively from the view of matrix were put forward. Furthermore, the idea of matrix was used to research and discuss the approaches for incrementally updating approximations of sets, based on the dynamic number of coverings. The investigations enriched and improved the covering rough set based dynamic learning theory and provided a method for dynamic knowledge update based in covering information systems.

To measure the pitch of twisted-pair wires, a kind of image detection framework was put forward. With image segmentation, image restoration, image thinning, curve fitting and scale setting, the pitch of twisted-pair wires was calculated in real time. In combination with this framework, to deal with the problem that the traditional two-dimensional maximum between-cluster variance algorithm (Otsu) runs too slow, a new fast algorithm based on regional diagonal points was proposed. With redefining two-dimensional histogram area, using the quick lookup table and recursion method, it reduced running time drastically. To solve the problem of image missing, an edge detection algorithm was adopted. After repairing, the image thinning operation was acted on the image. The least square method was used to fit the single pixel point of thinning image, then fitting curve was acquired. It could acquire the pitch of twisted-pair wires in the image by calculating the distance between the fitting curve intersections. Finally the distance in image was converted to an observed value by the scale. The experimental results show that the segmentation time of fast algorithm is about 0.22% of traditional algorithm. And two segmentation results of algorithms are identical. With the pitch from the image detection method comparing with its real value, results show that the absolute errors between both of them are 0.48%. Through the image detection method, the pitch is measured accurately and the efficiency of twisted-pair pitch measurement is improved.

An affine invariant local feature detector has been put forward for 3D object image matching. In order to cope with view angle and scale changes, this algorithm changed the image structure to fit the circular Gaussian filter according to the principle that Gaussian filter and image structure should be compatible. Local image structures were measured by covariance matrixes of Maximally Stable Extremal Regions (MSER) having been detected in the image. Anisotropic image structures must be rotated and squeezed into isotropic image structures to guarantee the correctness of image transformation. Finally, Scale Invariant Feature Transform (SIFT) features were extracted on isotropic image structures. Coordinates of SIFT features should be changed into the original image coordinates after being extracted. The experimental results indicate that the local features extracted by this algorithm are fully affine invariant. They are suitable to be used in wide baseline image matching for 3D objects.

A novel algorithm for mining user navigation pattern with incremental clustering was presented. Firstly, a new method for expressing user interest was introduced to construct user profile object. Based on the basic concept of ant colony clustering, artificial ants were used to pick up or drop down object to implement clustering by analyzing the similarity with other local regional objects and. Then a mechanism of decomposing clusters was used to form new clusters when users'interests changed. Experimental results show that the method can adaptively and efficiently achieve incremental clustering.