Long-distance small target detection:
In the technology of moving target detection, the false alarm has always been a difficult problem, especially for small targets. Most manufacturers only have moving target detection and lack certain processing abilities for the false alarm. Individual enterprises can have certain anti-interference abilities for periodic moving targets and batch targets through algorithm optimization. Through multi-technology integration and scene algorithm optimization, our company can have a certain anti-interference ability for all kinds of small targets with false alarms, and the minimum of real targets and false alarm targets can reach 5x5 pixels.
Remote small target recognition:
At present, the vast majority of recognition algorithms on the market are for the recognition of close-range targets with a distance of more than 200x200 pixels and less than 100m. Targets with a distance of more than 500m can hardly achieve the function of classification and recognition. Relying on photoelectric imaging technology, our company mainly carries out long-distance target recognition of more than 500m, which can effectively recognize targets with 30x30 pixels, and deeply optimize the classification and recognition algorithm in special scenes. The optimized algorithm has a certain classification and recognition ability for ultra-small targets of 15x15.
Thermal imaging video algorithm:
In addition to the image algorithm of visible video, our algorithm also has deep research on the image algorithm of thermal imaging video. Compared with the image algorithm of visible video in most similar industries, our algorithm has a certain leading advantage.