In 2009, Xunshu launched the "Image based Plankton Detection and Intelligent Identification System", which catered to the country's emphasis on environmental monitoring and provided effective means for algae monitoring and research in environmental monitoring institutions and research institutes.
In the process of algae identification, it is generally based on the specific morphological characteristics of the algal body. Due to the complex morphological features of the algal body, the morphology displayed at different stages and angles varies, which poses significant challenges to the identification of algal species. In the original algae system, the morphological search based identification system of Xunshu *, combined with the algae expert database for search and identification, can effectively narrow down the range of algae species and reduce workload.
In the new generation algae counting system launched in 2014, two new "fast counting" core technologies were added: high-precision intelligent search for biological similarity and chaotic intelligent classification counting, achieving fast algae assisted identification and automatic segmentation and counting of different algae. High precision intelligent search for biological similarity is the core technology of Xunshu's new generation of intelligent identification of algae. Through the effective combination of "morphological similarity" and "biological similarity", the biological characteristics of algae species are accurately extracted and fused, greatly improving the accuracy of algae search and making rapid algae identification possible.

Principle:

1) Color feature extraction: Extract algal cell color features based on color histograms.

2) Texture feature extraction: Based on Gabor filter rotation invariant features, extract algal cell texture features.

3) Intelligent search: Fusing two types of features into a feature vector, using a support vector machine classifier for training, to achieve classification and identification search of algal cell images.

example:

Automatic classification and counting of mixed algae: Chaos intelligent classification and counting is a major technological breakthrough in the research of algae automatic classification and counting by Xunshu, which has initially achieved automatic classification and counting of multiple types of algae cells with large differences in morphology and color.

Principle:

1) Chaos Genetic Algorithm: By utilizing the randomness, ergodicity, and initial sensitivity of chaotic motion, the chaotic state is introduced into the optimization variable, expanding the traversal range of chaotic motion to the value range of the optimization variable, thereby achieving image segmentation of all algae in the water body.

2) Fuzzy C-means clustering algorithm: Determine the degree to which each algal species sample data belongs to a certain cluster, and group algae with similar degrees of membership into one cluster.

example: