<p>Signal analysis of the weight of moving objects within a sealed cavity is a convenient and effective diagnostic method for finding faults indirectly without opening the shell. However, qualitative evaluation of weight remains a challenging problem due to the insufficient information available to effectively characterize different weights. To comprehensively assess different weight levels of moving objects from various perspectives, this study proposes a qualitative evaluation method for the weight of moving objects within a sealed cavity based on time-frequency spectrogram features. In this research, the impact of different weight levels on time-frequency spectrogram construction was first investigated. Then, features representing weight information were effectively integrated from the image, time-domain, frequency-domain, and time-frequency-domain aspects, bringing the qualitative evaluation of weight to new heights. It was also demonstrated that image features can be utilized to represent weight information effectively.</p>